d6520673-0f5f-4c28-898b-f52d056b257d.txt

import os
import sys
with open(sys.argv[0]) as f:
    code = f.read() # read the code of this file ASAP, for logging
import uuid
import glob
import time
import contextlib
from dataclasses import dataclass

import numpy as np
import torch
from torch import nn
import torch.nn.functional as F
import torch.distributed as dist
import torch._inductor.config as config
from torch.nn.parallel import DistributedDataParallel as DDP
# Use of FlexAttention contributed by @KoszarskyB
from torch.nn.attention.flex_attention import flex_attention, create_block_mask
flex_attention = torch.compile(flex_attention, dynamic=False)
create_block_mask = torch.compile(create_block_mask, dynamic=False)

# -----------------------------------------------------------------------------
# Muon optimizer

def zeropower_via_svd(G, steps=None):
    U, S, V = G.svd()
    return U @ V.T

@torch.compile
def zeropower_via_newtonschulz5(G, steps=10, eps=1e-7):
    """
    Newton-Schulz iteration to compute the zeroth power / orthogonalization of G. We opt to use a
    quintic iteration whose coefficients are selected to maximize the slope at zero. For the purpose
    of minimizing steps, it turns out to be empirically effective to keep increasing the slope at
    zero even beyond the point where the iteration no longer converges all the way to one everywhere
    on the interval. This iteration therefore does not produce UV^T but rather something like US'V^T
    where S' is diagonal with S_{ii}' ~ Uniform(0.5, 1.5), which turns out not to hurt model
    performance at all relative to UV^T, where USV^T = G is the SVD.
    """
    assert len(G.shape) == 2
    a, b, c = (3.4445, -4.7750,  2.0315)
    X = G.bfloat16()
    X /= (X.norm() + eps) # ensure top singular value <= 1
    if G.size(0) > G.size(1):
        X = X.T
    for _ in range(steps):
        A = X @ X.T
        B = b * A + c * A @ A # adapted from suggestion by @jxbz, @leloykun, and @YouJiacheng
        X = a * X + B @ X
    if G.size(0) > G.size(1):
        X = X.T
    return X

zeropower_backends = dict(svd=zeropower_via_svd, newtonschulz5=zeropower_via_newtonschulz5)

class Muon(torch.optim.Optimizer):
    """
    Muon - MomentUm Orthogonalized by Newton-schulz

    Muon internally runs standard SGD-momentum, and then performs an orthogonalization post-
    processing step, in which each 2D parameter's update is replaced with the nearest orthogonal
    matrix. To efficiently orthogonalize each update, we use a Newton-Schulz iteration, which has
    the advantage that it can be stably run in bfloat16 on the GPU.

    Some warnings:
    - This optimizer assumes that all parameters passed in are 2D.
    - It should not be used for the embedding layer, the final fully connected layer, or any {0,1}-D
    parameters; those should all be optimized by a standard method (e.g., AdamW).
    - To use it with 4D convolutional filters, it works well to just flatten their last 3 dimensions.
    - We believe it is unlikely to work well for training with small batch size.
    - We believe it may not work well for finetuning pretrained models, but we haven't tested this.
    - We have not yet tried this optimizer for training scenarios larger than NanoGPT (124M).

    Arguments:
        lr: The learning rate used by the internal SGD.
        momentum: The momentum used by the internal SGD.
        nesterov: Whether to use Nesterov-style momentum in the internal SGD. (recommended)
        backend: The chosen backend for the orthogonalization step. (recommended: 'newtonschulz5')
        backend_steps: The number of iteration steps to use in the backend, if it is iterative.
    """
    def __init__(self, params, lr=0.02, momentum=0.95, nesterov=True,
                 backend='newtonschulz5', backend_steps=5):
        defaults = dict(lr=lr, momentum=momentum, nesterov=nesterov, backend=backend, backend_steps=backend_steps)
        super().__init__(params, defaults)

    def step(self):

        for group in self.param_groups:

            lr = group['lr']
            momentum = group['momentum']
            zeropower_backend = zeropower_backends[group['backend']]

            # generate weight updates in distributed fashion
            total_params = sum(p.numel() for p in group['params'])
            updates_flat = torch.zeros(total_params, device='cuda', dtype=torch.bfloat16)
            curr_idx = 0
            for i, p in enumerate(group['params']):
                # luckily this will perfectly distribute a transformer with multiple of 4 layers to 8 GPUs
                if i % int(os.environ['WORLD_SIZE']) == int(os.environ['RANK']):
                    g = p.grad
                    assert g is not None
                    state = self.state[p]
                    if 'momentum_buffer' not in state:
                        state['momentum_buffer'] = torch.zeros_like(g)
                    buf = state['momentum_buffer']
                    buf.mul_(momentum).add_(g)
                    g = g.add(buf, alpha=momentum) if group['nesterov'] else buf
                    g = zeropower_backend(g, steps=group['backend_steps'])
                    g *= max(1, g.size(0)/g.size(1))**0.5
                    updates_flat[curr_idx:curr_idx+p.numel()] = g.flatten()
                curr_idx += p.numel()

            # sync updates across devices. we are not memory-constrained so can do this simple deserialization
            dist.all_reduce(updates_flat, op=dist.ReduceOp.SUM)

            # deserialize and apply updates
            curr_idx = 0
            for p in group['params']:
                g = updates_flat[curr_idx:curr_idx+p.numel()].view_as(p.data).type_as(p.data)
                p.data.add_(g, alpha=-lr)
                curr_idx += p.numel()

# -----------------------------------------------------------------------------
# PyTorch nn.Module definitions for the GPT-2 model

def norm(x):
    return F.rms_norm(x, (x.size(-1),))

class CastedLinear(nn.Linear):

    def __init__(self, in_features, out_features):
        super().__init__(in_features, out_features, bias=False)

    def forward(self, x):
        return F.linear(x, self.weight.to(x.dtype))

class Rotary(torch.nn.Module):

    def __init__(self, dim, base=10000):
        super().__init__()
        self.register_buffer('inv_freq', (1 / base) ** (torch.arange(0, dim, 2) / dim))
        self.seq_len_cached = None
        self.cos_cached = None
        self.sin_cached = None

    def forward(self, x):
        seq_len = x.shape[1]
        if seq_len != self.seq_len_cached:
            t = torch.arange(seq_len, device=x.device)
            freqs = torch.outer(t, self.inv_freq)
            self.seq_len_cached = seq_len
            self.cos_cached = freqs.cos()
            self.sin_cached = freqs.sin()
        cos, sin = self.cos_cached[None, :, None, :], self.sin_cached[None, :, None, :]
        # apply_rotary_emb(x, cos, sin)
        x1, x2 = x.chunk(2, dim=3)
        y1 = x1 * cos + x2 * sin
        y2 = x1 * (-sin) + x2 * cos
        return torch.cat((y1, y2), 3).type_as(x)

class CausalSelfAttention(nn.Module):

    def __init__(self, dim, n_head):
        super().__init__()
        assert dim % n_head == 0
        self.n_head = n_head
        self.c_q = CastedLinear(dim, dim)
        self.c_k = CastedLinear(dim, dim)
        self.c_v = CastedLinear(dim, dim)
        # value residual lambda
        self.lamb = nn.Parameter(torch.tensor(0.5)) # @Grad62304977
        # rotary embeddings
        self.rotary = Rotary(dim // n_head) # dim // n_head = head_dim
        # output projection
        self.c_proj = CastedLinear(dim, dim)
        self.c_proj.weight.data.zero_() # zero init suggested by @Grad62304977

    def forward(self, x, vi, block_mask):
        B, T = x.size(0), x.size(1) # batch size, sequence length
        assert B == 1, "Must use batch size = 1 for FlexAttention"
        q = self.c_q(x).view(B, T, self.n_head, -1)
        k = self.c_k(x).view(B, T, self.n_head, -1)
        v = self.c_v(x).view(B, T, self.n_head, -1)
        v = (1 - self.lamb) * v + self.lamb * vi.view_as(v) # @Grad62304977
        q, k = norm(q), norm(k) # QK norm suggested by @Grad62304977
        q, k = self.rotary(q), self.rotary(k)
        y = flex_attention(q.transpose(1, 2), k.transpose(1, 2), v.transpose(1, 2), block_mask=block_mask)
        y = y.transpose(1, 2).contiguous().view_as(x) # re-assemble all head outputs side by side
        y = self.c_proj(y)
        return y

class MLP(nn.Module):

    def __init__(self, dim):
        super().__init__()
        self.c_fc   = CastedLinear(dim, 4 * dim)
        self.c_proj = CastedLinear(4 * dim, dim)
        self.c_proj.weight.data.zero_() # zero init suggested by @Grad62304977

    def forward(self, x):
        x = self.c_fc(x)
        x = F.relu(x).square() # https://arxiv.org/abs/2109.08668v2; ~1-2% better than GELU; suggested by @SKYLINEZ007 and @Grad62304977
        x = self.c_proj(x)
        return x

class Block(nn.Module):

    def __init__(self, config):
        super().__init__()
        self.attn = CausalSelfAttention(config.n_embd, config.n_head)
        self.mlp = MLP(config.n_embd)
        self.lambdas = nn.Parameter(torch.tensor([1., 0.]))

    def forward(self, x, vi, x0, block_mask):
        x = self.lambdas[0] * x + self.lambdas[1] * x0
        x = x + self.attn(norm(x), vi, block_mask)
        x = x + self.mlp(norm(x))
        return x

# -----------------------------------------------------------------------------
# The main GPT-2 model

@dataclass
class GPTConfig:
    vocab_size : int = 50304
    n_layer : int = 12
    n_head : int = 6 # head dim 128 suggested by @Grad62304977
    n_embd : int = 768

class GPT(nn.Module):

    def __init__(self, config):
        super().__init__()

        # U-net design by @brendanh0gan
        self.num_encoder_layers = config.n_layer // 2 # Half of the layers for encoder
        self.num_decoder_layers = config.n_layer - self.num_encoder_layers # Remaining for decoder
        # Add learnable skip connection weights for decoder layers
        self.skip_weights = nn.Parameter(torch.ones(self.num_decoder_layers))

        self.transformer = nn.ModuleDict(dict(
            wte = nn.Embedding(config.vocab_size, config.n_embd),
            # token value embeddings by @KoszarskyB - inspired by @Grad62304977's value residual learning
            vte = nn.Embedding(config.vocab_size, config.n_embd*12),
            h = nn.ModuleList([Block(config) for _ in range(config.n_layer)]),
        ))
        self.lm_head = CastedLinear(config.n_embd, config.vocab_size)
        self.lm_head.weight.data.zero_() # @Grad62304977

    def forward(self, idx, target, attn_blocksize):

        docs = (idx == 50256).cumsum(0)
        def document_causal_mask(b, h, q_idx, kv_idx):
          causal_mask = q_idx >= kv_idx
          document_mask = docs[q_idx] == docs[kv_idx]
          window_mask = q_idx - kv_idx < attn_blocksize
          return causal_mask & document_mask & window_mask

        S = len(idx)
        block_mask = create_block_mask(document_causal_mask, None, None, S, S, device="cuda", _compile=True)

        # forward the GPT model itself
        x = self.transformer.wte(idx[None]) # token embeddings of shape (b, t, n_embd)
        x = norm(x) # @Grad62304977
        x0 = x
        vi = self.transformer.vte(idx[None]).chunk(12, dim=-1)

        # Store outputs for U-Net skip connections
        skip_connections = []
        # Encoder pass - process only the first half of the blocks
        for i in range(self.num_encoder_layers):
            x = self.transformer.h[i](x, vi[i], x0, block_mask)
            skip_connections.append(x)
        # Decoder pass - process the remaining blocks with weighted skip connections
        for i in range(self.num_decoder_layers):
            x = x + self.skip_weights[i] * skip_connections.pop()
            x = self.transformer.h[self.num_encoder_layers + i](x, vi[self.num_encoder_layers+i], x0, block_mask)

        x = norm(x)
        logits = self.lm_head(x)
        logits = 30 * torch.tanh(logits / 30) # @Grad62304977
        logits = logits.float()
        loss = F.cross_entropy(logits.view(-1, logits.size(-1)), target.view(-1))
        return loss

# -----------------------------------------------------------------------------
# Our own simple Distributed Data Loader

def _peek_data_shard(filename):
    # only reads the header, returns header data
    with open(filename, "rb") as f:
        # first read the header, which is 256 int32 integers (4 bytes each)
        header = np.frombuffer(f.read(256*4), dtype=np.int32)
    if header[0] != 20240520:
        print("ERROR: magic number mismatch in the data .bin file!")
        print("---> HINT: Are you passing in a correct file with --input_bin?")
        print("---> HINT: Dataset encoding changed recently, re-run data prepro or refer again to README")
        print("---> HINT: For example re-run: `python dev/data/tinyshakespeare.py`, then re-try")
        exit(1)
    assert header[1] == 1, "unsupported version"
    ntok = header[2] # number of tokens (claimed)
    return ntok # for now just return the number of tokens

def _load_data_shard(filename):
    with open(filename, "rb") as f:
        # first read the header, which is 256 int32 integers (4 bytes each)
        header = np.frombuffer(f.read(256*4), dtype=np.int32)
        assert header[0] == 20240520, "magic number mismatch in the data .bin file"
        assert header[1] == 1, "unsupported version"
        ntok = header[2] # number of tokens (claimed)
        # the rest of it are tokens, stored as uint16
        tokens = np.frombuffer(f.read(), dtype=np.uint16)
    assert len(tokens) == ntok, "number of tokens read does not match header?"
    return tokens

class DistributedDataLoader:
    def __init__(self, filename_pattern, T, process_rank, num_processes):
        self.process_rank = process_rank
        self.num_processes = num_processes
        self.T = T

        # glob files that match the pattern
        self.files = sorted(glob.glob(filename_pattern))
        assert len(self.files) > 0, f"did not find any files that match the pattern {filename_pattern}"

        # load and validate all data shards, count number of tokens in total
        ntok_total = 0
        for fname in self.files:
            shard_ntok = _peek_data_shard(fname)
            assert shard_ntok >= num_processes * T + 1
            ntok_total += int(shard_ntok)
        self.ntok_total = ntok_total

        self.reset()

    def reset(self):
        self.current_shard = -1
        self.advance()

    def advance(self): # advance to next data shard
        self.current_shard = (self.current_shard + 1) % len(self.files)
        self.current_position = self.process_rank * self.T
        self.tokens = _load_data_shard(self.files[self.current_shard])

    def next_batch(self):
        batch_size = self.T * self.num_processes
        buf = self.tokens[self.current_position:self.current_position+self.T+1]
        buf = torch.tensor(buf.astype(np.int32), dtype=torch.long)
        x = buf[:-1] # inputs
        y = buf[1:] # targets
        # advance current position and load next shard if necessary
        self.current_position += batch_size
        if self.current_position + batch_size >= len(self.tokens):
            self.advance()
        return x.cuda(), y.cuda()

# -----------------------------------------------------------------------------
# int main

@dataclass
class Hyperparameters:
    # data hyperparams
    input_bin : str = 'data/fineweb10B/fineweb_train_*.bin' # input .bin to train on
    input_val_bin : str = 'data/fineweb10B/fineweb_val_*.bin' # input .bin to eval validation loss on
    # optimization hyperparams
    batch_size : int = 8 # batch size, in sequences, across all devices
    sequence_length : int = 64*1024 # sequence length, in tokens
    num_iterations : int = 1530 # number of iterations to run
    warmup_iters : int = 0
    cooldown_iters : int = 600 # number of iterations of linear warmup/cooldown for triangular or trapezoidal schedule
    weight_decay : float = 0
    # evaluation and logging hyperparams
    val_loss_every : int = 125 # every how many steps to evaluate val loss? 0 for only at the end
    val_tokens : int = 10485760 # how many tokens of validation data? it's important to keep this fixed for consistent comparisons
    save_every : int = 0 # every how many steps to save the checkpoint? 0 for only at the end
args = Hyperparameters()

# set up DDP (distributed data parallel). torchrun sets this env variable
assert torch.cuda.is_available()
dist.init_process_group(backend='nccl')
ddp_rank = int(os.environ['RANK'])
ddp_local_rank = int(os.environ['LOCAL_RANK'])
ddp_world_size = int(os.environ['WORLD_SIZE'])
device = f'cuda:{ddp_local_rank}'
torch.cuda.set_device(device)
print(f"using device: {device}")
master_process = (ddp_rank == 0) # this process will do logging, checkpointing etc.

# begin logging
logfile = None
if master_process:
    run_id = str(uuid.uuid4())
    logdir = 'logs/%s/' % run_id
    os.makedirs(logdir, exist_ok=True)
    logfile = 'logs/%s.txt' % run_id
    # create the log file
    with open(logfile, "w") as f:
        # begin the log by printing this file (the Python code)
        f.write(code)
        f.write('='*100 + '\n')
def print0(s, logonly=False):
    if master_process:
        with open(logfile, "a") as f:
            if not logonly:
                print(s)
            f.write(s+'\n')
# log information about the hardware/software environment this is running on
# and print the full `nvidia-smi` to file
print0(f"Running pytorch {torch.version.__version__} compiled for CUDA {torch.version.cuda}\nnvidia-smi:")
import subprocess
result = subprocess.run(['nvidia-smi'], stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True)
print0(f'{result.stdout}', logonly=True)
print0('='*100, logonly=True)

# convenience variables
T = args.sequence_length
# calculate the number of steps to take in the val loop.
assert args.val_tokens % (T * ddp_world_size) == 0
val_steps = args.val_tokens // (T * ddp_world_size)
# calculate the steps of gradient accumulation required to attain the desired global batch size.
assert args.batch_size % (ddp_world_size) == 0
train_accumulation_steps = args.batch_size // ddp_world_size

# load tokens
train_loader = DistributedDataLoader(args.input_bin, T, ddp_rank, ddp_world_size)
val_loader = DistributedDataLoader(args.input_val_bin, T, ddp_rank, ddp_world_size)
print0(f"Training DataLoader: total number of tokens: {train_loader.ntok_total} across {len(train_loader.files)} files")
print0(f"Validation DataLoader: total number of tokens: {val_loader.ntok_total} across {len(val_loader.files)} files")
print0('='*100, logonly=True)
x, y = train_loader.next_batch()

# there are only 50257 unique GPT-2 tokens; we extend to nearest multiple of 128 for efficiency. suggested to me by @Grad62304977.
# this originates from Karpathy's experiments.
num_vocab = 50304
model = GPT(GPTConfig(vocab_size=num_vocab, n_layer=12, n_head=6, n_embd=768))
model = model.cuda().bfloat16()
for m in model.modules():
    if isinstance(m, CastedLinear):
        m.float()
if hasattr(config, "coordinate_descent_tuning"):
    config.coordinate_descent_tuning = True # suggested by @Chillee
model = torch.compile(model)
# here we wrap model into DDP container
model = DDP(model, device_ids=[ddp_local_rank])
raw_model = model.module # always contains the "raw" unwrapped model

# init the optimizer(s)
optimizer1 = torch.optim.Adam([raw_model.transformer.wte.weight, raw_model.transformer.vte.weight], lr=0.6, betas=(0.8, 0.95), fused=True)
optimizer2 = torch.optim.Adam([raw_model.lm_head.weight], lr=0.008, betas=(0.8, 0.95), fused=True)
params = list(raw_model.transformer.h.parameters())
matrix_params = [p for p in params if p.ndim == 2]
scalar_params = [p for p in params if p.ndim < 2] + [raw_model.skip_weights]
optimizer3 = Muon(matrix_params, lr=0.05, momentum=0.95)
optimizer4 = torch.optim.Adam(scalar_params, lr=0.04, betas=(0.8, 0.95), fused=True) # note that this learning rate is neither sensitive nor tuned
optimizers = [optimizer1, optimizer2, optimizer3, optimizer4]
# learning rate decay scheduler (linear warmup and cooldown)
def get_lr(it):
    assert it <= args.num_iterations
    # 1) linear warmup for warmup_iters steps
    if it < args.warmup_iters:
        return (it+1) / args.warmup_iters
    # 2) constant lr for a while
    elif it < args.num_iterations - args.cooldown_iters:
        return 1.0
    # 3) linear cooldown
    else:
        decay_ratio = (args.num_iterations - it) / args.cooldown_iters
        return decay_ratio
schedulers = [torch.optim.lr_scheduler.LambdaLR(opt, get_lr) for opt in optimizers]

# Start training loop
training_time_ms = 0
# start the clock
torch.cuda.synchronize()
t0 = time.time()
# begin training
for step in range(args.num_iterations + 1):
    last_step = (step == args.num_iterations)
    # This effectively ignores timing first 10 steps, which are slower for weird reasons.
    # Alternately, and slightly more correctly in terms of benchmarking, we could do 10
    # steps with dummy data first, and then re-initialize the model and reset the loader.
    if step == 10:
        training_time_ms = 0
        t0 = time.time()
    timed_steps = float('nan') if step <= 11 else (step - 10) + 1 # <= 11 to avoid bug in val

    # Set the attention blocksize for the current step, in chunks of 64. By @fernbear.bsky.social
    attn_blocksize = torch.tensor(64*((step/args.num_iterations * (1792 - 64) + 64)//64), dtype=torch.int, device='cuda')

    # once in a while evaluate the validation dataset
    if (last_step or (args.val_loss_every > 0 and step % args.val_loss_every == 0)):
        # stop the clock
        torch.cuda.synchronize()
        training_time_ms += 1000 * (time.time() - t0)
        # run validation batches
        model.eval()
        val_loader.reset()
        val_loss = 0.0
        for _ in range(val_steps):
            with torch.no_grad():
                x_val, y_val = val_loader.next_batch()
                val_loss += model(x_val, y_val, attn_blocksize=attn_blocksize)
        dist.all_reduce(val_loss, op=dist.ReduceOp.AVG)
        val_loss /= val_steps
        # log val loss to console and to logfile
        print0(f'step:{step}/{args.num_iterations} val_loss:{val_loss:.4f} train_time:{training_time_ms:.0f}ms step_avg:{training_time_ms/(timed_steps-1):.2f}ms')
        # start the clock again
        torch.cuda.synchronize()
        t0 = time.time()

    if master_process and (last_step or (args.save_every > 0 and step % args.save_every == 0)):
        # stop the clock
        torch.cuda.synchronize()
        training_time_ms += 1000 * (time.time() - t0)
        # save the state of the training process
        log = dict(step=step, code=code, model=raw_model.state_dict(), optimizers=[opt.state_dict() for opt in optimizers])
        torch.save(log, 'logs/%s/state_step%06d.pt' % (run_id, step))
        # start the clock again
        torch.cuda.synchronize()
        t0 = time.time()

    # bit confusing: we want to make sure to eval on 0th iteration
    # but also after the very last iteration. so we loop for step <= num_iterations
    # instead of just < num_iterations (one extra due to <=), only to do
    # the validation/sampling one last time, and then we break right here as we're done.
    if last_step:
        break

    # --------------- TRAINING SECTION BEGIN -----------------
    model.train()
    for i in range(1, train_accumulation_steps+1):
        ctx = model.no_sync() if i < train_accumulation_steps else contextlib.nullcontext()
        with ctx: # there's no need to sync gradients every accumulation step
            # forward pass
            loss = model(x, y, attn_blocksize=attn_blocksize)
            # advance the dataset for the next batch
            x, y = train_loader.next_batch()
            # backward pass
            loss.backward()
        train_loss = loss.detach()
    for p in model.parameters():
        p.grad /= train_accumulation_steps
    # momentum warmup for Muon
    frac = min(step/300, 1)
    optimizer3.param_groups[0]['momentum'] = (1 - frac) * 0.85 + frac * 0.95
    # step the optimizers and schedulers
    for opt, sched in zip(optimizers, schedulers):
        opt.step()
        sched.step()
    # null the gradients
    model.zero_grad(set_to_none=True)
    # --------------- TRAINING SECTION END -------------------
    # everything that follows now is just diagnostics, prints, logging, etc.

    #dist.all_reduce(train_loss, op=dist.ReduceOp.AVG) # all-reducing the training loss would be more correct in terms of logging, but slower
    approx_time = training_time_ms + 1000 * (time.time() - t0)
    print0(f"step:{step+1}/{args.num_iterations} train_loss:{train_loss.item():.4f} train_time:{approx_time:.0f}ms step_avg:{approx_time/timed_steps:.2f}ms")

if master_process:
    print(f"peak memory consumption: {torch.cuda.max_memory_allocated() // 1024 // 1024} MiB")

# -------------------------------------------------------------------------
# clean up nice
dist.destroy_process_group()
====================================================================================================
Running pytorch 2.6.0.dev20241203+cu124 compiled for CUDA 12.4
nvidia-smi:
Thu Dec  5 02:19:36 2024       
+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 535.183.06             Driver Version: 535.183.06   CUDA Version: 12.2     |
|-----------------------------------------+----------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |         Memory-Usage | GPU-Util  Compute M. |
|                                         |                      |               MIG M. |
|=========================================+======================+======================|
|   0  NVIDIA H100 80GB HBM3          On  | 00000000:19:00.0 Off |                    0 |
| N/A   38C    P0              75W / 700W |      3MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   1  NVIDIA H100 80GB HBM3          On  | 00000000:3B:00.0 Off |                    0 |
| N/A   30C    P0             115W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   2  NVIDIA H100 80GB HBM3          On  | 00000000:4C:00.0 Off |                    0 |
| N/A   31C    P0             118W / 700W |    529MiB / 81559MiB |      1%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   3  NVIDIA H100 80GB HBM3          On  | 00000000:5D:00.0 Off |                    0 |
| N/A   38C    P0             118W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   4  NVIDIA H100 80GB HBM3          On  | 00000000:9B:00.0 Off |                    0 |
| N/A   39C    P0             123W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   5  NVIDIA H100 80GB HBM3          On  | 00000000:BB:00.0 Off |                    0 |
| N/A   29C    P0             110W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   6  NVIDIA H100 80GB HBM3          On  | 00000000:CB:00.0 Off |                    0 |
| N/A   39C    P0             127W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   7  NVIDIA H100 80GB HBM3          On  | 00000000:DB:00.0 Off |                    0 |
| N/A   30C    P0             118W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
                                                                                         
+---------------------------------------------------------------------------------------+
| Processes:                                                                            |
|  GPU   GI   CI        PID   Type   Process name                            GPU Memory |
|        ID   ID                                                             Usage      |
|=======================================================================================|
+---------------------------------------------------------------------------------------+

====================================================================================================
Training DataLoader: total number of tokens: 1100000000 across 11 files
Validation DataLoader: total number of tokens: 100000000 across 1 files
====================================================================================================
step:0/1530 val_loss:10.8258 train_time:0ms step_avg:nanms
step:1/1530 train_loss:10.8258 train_time:31485ms step_avg:nanms
step:2/1530 train_loss:10.0780 train_time:31596ms step_avg:nanms
step:3/1530 train_loss:8.3799 train_time:31755ms step_avg:nanms
step:4/1530 train_loss:7.5771 train_time:31916ms step_avg:nanms
step:5/1530 train_loss:7.4605 train_time:32077ms step_avg:nanms
step:6/1530 train_loss:6.9770 train_time:32237ms step_avg:nanms
step:7/1530 train_loss:7.1970 train_time:32399ms step_avg:nanms
step:8/1530 train_loss:6.7437 train_time:32559ms step_avg:nanms
step:9/1530 train_loss:6.6149 train_time:32719ms step_avg:nanms
step:10/1530 train_loss:6.4913 train_time:32879ms step_avg:nanms
step:11/1530 train_loss:6.4531 train_time:114ms step_avg:nanms
step:12/1530 train_loss:6.3716 train_time:276ms step_avg:nanms
step:13/1530 train_loss:6.2501 train_time:436ms step_avg:145.23ms
step:14/1530 train_loss:6.2664 train_time:596ms step_avg:148.92ms
step:15/1530 train_loss:6.1678 train_time:756ms step_avg:151.28ms
step:16/1530 train_loss:6.1338 train_time:917ms step_avg:152.77ms
step:17/1530 train_loss:6.1632 train_time:1075ms step_avg:153.64ms
step:18/1530 train_loss:5.9654 train_time:1237ms step_avg:154.60ms
step:19/1530 train_loss:5.9700 train_time:1397ms step_avg:155.20ms
step:20/1530 train_loss:5.7142 train_time:1556ms step_avg:155.61ms
step:21/1530 train_loss:5.9572 train_time:1717ms step_avg:156.13ms
step:22/1530 train_loss:6.1560 train_time:1878ms step_avg:156.51ms
step:23/1530 train_loss:5.8443 train_time:2038ms step_avg:156.80ms
step:24/1530 train_loss:6.0306 train_time:2198ms step_avg:156.99ms
step:25/1530 train_loss:5.6815 train_time:2359ms step_avg:157.25ms
step:26/1530 train_loss:5.5955 train_time:2520ms step_avg:157.48ms
step:27/1530 train_loss:5.7527 train_time:2679ms step_avg:157.61ms
step:28/1530 train_loss:5.4114 train_time:2841ms step_avg:157.82ms
step:29/1530 train_loss:5.6747 train_time:3001ms step_avg:157.93ms
step:30/1530 train_loss:5.4714 train_time:3162ms step_avg:158.09ms
step:31/1530 train_loss:5.4282 train_time:3322ms step_avg:158.20ms
step:32/1530 train_loss:5.2931 train_time:3484ms step_avg:158.38ms
step:33/1530 train_loss:5.5779 train_time:3643ms step_avg:158.41ms
step:34/1530 train_loss:5.4965 train_time:3803ms step_avg:158.47ms
step:35/1530 train_loss:5.6091 train_time:3964ms step_avg:158.57ms
step:36/1530 train_loss:5.5537 train_time:4125ms step_avg:158.64ms
step:37/1530 train_loss:5.4458 train_time:4286ms step_avg:158.74ms
step:38/1530 train_loss:5.3028 train_time:4447ms step_avg:158.82ms
step:39/1530 train_loss:5.3212 train_time:4608ms step_avg:158.88ms
step:40/1530 train_loss:5.2566 train_time:4768ms step_avg:158.93ms
step:41/1530 train_loss:5.2313 train_time:4928ms step_avg:158.98ms
step:42/1530 train_loss:5.1684 train_time:5088ms step_avg:159.01ms
step:43/1530 train_loss:5.2618 train_time:5248ms step_avg:159.04ms
step:44/1530 train_loss:5.2456 train_time:5408ms step_avg:159.07ms
step:45/1530 train_loss:5.3883 train_time:5569ms step_avg:159.12ms
step:46/1530 train_loss:5.1657 train_time:5729ms step_avg:159.13ms
step:47/1530 train_loss:5.0603 train_time:5889ms step_avg:159.17ms
step:48/1530 train_loss:5.2192 train_time:6049ms step_avg:159.19ms
step:49/1530 train_loss:5.1393 train_time:6209ms step_avg:159.21ms
step:50/1530 train_loss:5.2391 train_time:6370ms step_avg:159.25ms
step:51/1530 train_loss:5.1286 train_time:6530ms step_avg:159.26ms
step:52/1530 train_loss:5.0282 train_time:6689ms step_avg:159.27ms
step:53/1530 train_loss:5.1771 train_time:6849ms step_avg:159.29ms
step:54/1530 train_loss:5.0327 train_time:7010ms step_avg:159.31ms
step:55/1530 train_loss:5.4132 train_time:7170ms step_avg:159.34ms
step:56/1530 train_loss:5.0209 train_time:7330ms step_avg:159.34ms
step:57/1530 train_loss:4.8637 train_time:7490ms step_avg:159.36ms
step:58/1530 train_loss:5.0346 train_time:7650ms step_avg:159.37ms
step:59/1530 train_loss:5.0077 train_time:7811ms step_avg:159.40ms
step:60/1530 train_loss:5.1231 train_time:7971ms step_avg:159.41ms
step:61/1530 train_loss:4.8402 train_time:8130ms step_avg:159.42ms
step:62/1530 train_loss:4.9702 train_time:8290ms step_avg:159.42ms
step:63/1530 train_loss:4.9658 train_time:8450ms step_avg:159.44ms
step:64/1530 train_loss:4.9399 train_time:8610ms step_avg:159.44ms
step:65/1530 train_loss:4.7817 train_time:8770ms step_avg:159.45ms
step:66/1530 train_loss:4.9064 train_time:8930ms step_avg:159.46ms
step:67/1530 train_loss:4.8141 train_time:9090ms step_avg:159.47ms
step:68/1530 train_loss:5.1025 train_time:9250ms step_avg:159.48ms
step:69/1530 train_loss:4.7342 train_time:9411ms step_avg:159.51ms
step:70/1530 train_loss:4.8741 train_time:9571ms step_avg:159.52ms
step:71/1530 train_loss:4.9626 train_time:9731ms step_avg:159.52ms
step:72/1530 train_loss:4.8734 train_time:9890ms step_avg:159.52ms
step:73/1530 train_loss:4.7499 train_time:10051ms step_avg:159.54ms
step:74/1530 train_loss:4.8983 train_time:10211ms step_avg:159.54ms
step:75/1530 train_loss:4.8382 train_time:10371ms step_avg:159.56ms
step:76/1530 train_loss:4.7848 train_time:10532ms step_avg:159.57ms
step:77/1530 train_loss:4.9067 train_time:10692ms step_avg:159.58ms
step:78/1530 train_loss:5.1151 train_time:10852ms step_avg:159.58ms
step:79/1530 train_loss:4.7755 train_time:11011ms step_avg:159.58ms
step:80/1530 train_loss:4.8312 train_time:11174ms step_avg:159.63ms
step:81/1530 train_loss:4.6272 train_time:11335ms step_avg:159.65ms
step:82/1530 train_loss:4.8113 train_time:11495ms step_avg:159.66ms
step:83/1530 train_loss:4.7512 train_time:11657ms step_avg:159.68ms
step:84/1530 train_loss:4.7470 train_time:11817ms step_avg:159.69ms
step:85/1530 train_loss:4.6131 train_time:11978ms step_avg:159.71ms
step:86/1530 train_loss:4.8217 train_time:12137ms step_avg:159.70ms
step:87/1530 train_loss:4.7306 train_time:12298ms step_avg:159.72ms
step:88/1530 train_loss:4.7478 train_time:12459ms step_avg:159.73ms
step:89/1530 train_loss:4.6856 train_time:12620ms step_avg:159.74ms
step:90/1530 train_loss:4.6290 train_time:12781ms step_avg:159.76ms
step:91/1530 train_loss:4.6200 train_time:12941ms step_avg:159.77ms
step:92/1530 train_loss:4.7790 train_time:13102ms step_avg:159.78ms
step:93/1530 train_loss:4.5871 train_time:13262ms step_avg:159.79ms
step:94/1530 train_loss:4.6217 train_time:13422ms step_avg:159.79ms
step:95/1530 train_loss:4.6599 train_time:13585ms step_avg:159.82ms
step:96/1530 train_loss:4.5822 train_time:13744ms step_avg:159.81ms
step:97/1530 train_loss:4.6175 train_time:13904ms step_avg:159.82ms
step:98/1530 train_loss:4.5844 train_time:14065ms step_avg:159.83ms
step:99/1530 train_loss:4.6472 train_time:14224ms step_avg:159.82ms
step:100/1530 train_loss:4.6633 train_time:14386ms step_avg:159.84ms
step:101/1530 train_loss:4.5219 train_time:14546ms step_avg:159.85ms
step:102/1530 train_loss:4.6828 train_time:14707ms step_avg:159.86ms
step:103/1530 train_loss:4.5541 train_time:14868ms step_avg:159.87ms
step:104/1530 train_loss:4.5388 train_time:15028ms step_avg:159.87ms
step:105/1530 train_loss:4.5397 train_time:15188ms step_avg:159.87ms
step:106/1530 train_loss:4.5801 train_time:15348ms step_avg:159.88ms
step:107/1530 train_loss:4.4976 train_time:15509ms step_avg:159.88ms
step:108/1530 train_loss:4.3599 train_time:15669ms step_avg:159.89ms
step:109/1530 train_loss:4.4850 train_time:15829ms step_avg:159.89ms
step:110/1530 train_loss:4.4688 train_time:15988ms step_avg:159.88ms
step:111/1530 train_loss:4.4056 train_time:16149ms step_avg:159.89ms
step:112/1530 train_loss:4.5775 train_time:16310ms step_avg:159.90ms
step:113/1530 train_loss:4.4786 train_time:16470ms step_avg:159.90ms
step:114/1530 train_loss:4.3497 train_time:16630ms step_avg:159.90ms
step:115/1530 train_loss:4.4838 train_time:16792ms step_avg:159.93ms
step:116/1530 train_loss:4.4512 train_time:16956ms step_avg:159.96ms
step:117/1530 train_loss:4.3605 train_time:17121ms step_avg:160.01ms
step:118/1530 train_loss:4.5897 train_time:17286ms step_avg:160.06ms
step:119/1530 train_loss:4.4455 train_time:17448ms step_avg:160.08ms
step:120/1530 train_loss:4.3207 train_time:17612ms step_avg:160.11ms
step:121/1530 train_loss:4.2856 train_time:17777ms step_avg:160.15ms
step:122/1530 train_loss:4.4393 train_time:17942ms step_avg:160.19ms
step:123/1530 train_loss:4.2840 train_time:18105ms step_avg:160.22ms
step:124/1530 train_loss:4.5766 train_time:18269ms step_avg:160.25ms
step:125/1530 train_loss:4.4369 train_time:18432ms step_avg:160.28ms
step:125/1530 val_loss:4.3846 train_time:18479ms step_avg:160.69ms
step:126/1530 train_loss:4.3967 train_time:18599ms step_avg:160.34ms
step:127/1530 train_loss:4.4196 train_time:18767ms step_avg:160.40ms
step:128/1530 train_loss:4.3731 train_time:18930ms step_avg:160.42ms
step:129/1530 train_loss:4.6663 train_time:19095ms step_avg:160.46ms
step:130/1530 train_loss:4.3577 train_time:19261ms step_avg:160.51ms
step:131/1530 train_loss:4.4040 train_time:19425ms step_avg:160.54ms
step:132/1530 train_loss:4.3380 train_time:19588ms step_avg:160.56ms
step:133/1530 train_loss:4.4403 train_time:19753ms step_avg:160.59ms
step:134/1530 train_loss:4.2630 train_time:19916ms step_avg:160.61ms
step:135/1530 train_loss:4.4398 train_time:20080ms step_avg:160.64ms
step:136/1530 train_loss:4.2051 train_time:20244ms step_avg:160.67ms
step:137/1530 train_loss:4.3692 train_time:20408ms step_avg:160.69ms
step:138/1530 train_loss:4.2874 train_time:20571ms step_avg:160.71ms
step:139/1530 train_loss:4.3754 train_time:20736ms step_avg:160.74ms
step:140/1530 train_loss:4.4704 train_time:20901ms step_avg:160.78ms
step:141/1530 train_loss:4.3072 train_time:21066ms step_avg:160.81ms
step:142/1530 train_loss:4.3003 train_time:21229ms step_avg:160.83ms
step:143/1530 train_loss:4.2576 train_time:21393ms step_avg:160.85ms
step:144/1530 train_loss:4.3520 train_time:21556ms step_avg:160.86ms
step:145/1530 train_loss:4.2977 train_time:21720ms step_avg:160.89ms
step:146/1530 train_loss:4.1693 train_time:21884ms step_avg:160.91ms
step:147/1530 train_loss:4.3227 train_time:22047ms step_avg:160.93ms
step:148/1530 train_loss:4.3587 train_time:22210ms step_avg:160.94ms
step:149/1530 train_loss:4.2975 train_time:22373ms step_avg:160.96ms
step:150/1530 train_loss:4.4318 train_time:22536ms step_avg:160.97ms
step:151/1530 train_loss:4.2592 train_time:22702ms step_avg:161.00ms
step:152/1530 train_loss:4.2585 train_time:22866ms step_avg:161.03ms
step:153/1530 train_loss:4.3607 train_time:23029ms step_avg:161.04ms
step:154/1530 train_loss:4.3630 train_time:23194ms step_avg:161.07ms
step:155/1530 train_loss:4.2546 train_time:23358ms step_avg:161.09ms
step:156/1530 train_loss:4.3404 train_time:23522ms step_avg:161.11ms
step:157/1530 train_loss:4.4012 train_time:23685ms step_avg:161.12ms
step:158/1530 train_loss:4.2447 train_time:23849ms step_avg:161.14ms
step:159/1530 train_loss:4.2979 train_time:24012ms step_avg:161.15ms
step:160/1530 train_loss:4.1254 train_time:24175ms step_avg:161.17ms
step:161/1530 train_loss:4.3445 train_time:24340ms step_avg:161.19ms
step:162/1530 train_loss:4.3591 train_time:24504ms step_avg:161.21ms
step:163/1530 train_loss:4.3401 train_time:24669ms step_avg:161.24ms
step:164/1530 train_loss:4.1875 train_time:24833ms step_avg:161.26ms
step:165/1530 train_loss:4.2801 train_time:24998ms step_avg:161.28ms
step:166/1530 train_loss:4.3407 train_time:25163ms step_avg:161.30ms
step:167/1530 train_loss:4.1992 train_time:25326ms step_avg:161.31ms
step:168/1530 train_loss:4.2770 train_time:25490ms step_avg:161.33ms
step:169/1530 train_loss:4.1533 train_time:25655ms step_avg:161.35ms
step:170/1530 train_loss:4.0173 train_time:25819ms step_avg:161.37ms
step:171/1530 train_loss:4.1946 train_time:25981ms step_avg:161.37ms
step:172/1530 train_loss:4.2070 train_time:26144ms step_avg:161.39ms
step:173/1530 train_loss:4.2673 train_time:26307ms step_avg:161.39ms
step:174/1530 train_loss:4.4161 train_time:26469ms step_avg:161.40ms
step:175/1530 train_loss:4.2442 train_time:26633ms step_avg:161.41ms
step:176/1530 train_loss:4.0916 train_time:26796ms step_avg:161.42ms
step:177/1530 train_loss:4.0610 train_time:26960ms step_avg:161.44ms
step:178/1530 train_loss:4.1835 train_time:27123ms step_avg:161.45ms
step:179/1530 train_loss:4.1218 train_time:27286ms step_avg:161.45ms
step:180/1530 train_loss:4.1070 train_time:27449ms step_avg:161.46ms
step:181/1530 train_loss:4.2940 train_time:27611ms step_avg:161.47ms
step:182/1530 train_loss:4.1520 train_time:27773ms step_avg:161.47ms
step:183/1530 train_loss:4.1251 train_time:27938ms step_avg:161.49ms
step:184/1530 train_loss:4.1148 train_time:28101ms step_avg:161.50ms
step:185/1530 train_loss:4.2031 train_time:28264ms step_avg:161.51ms
step:186/1530 train_loss:4.1682 train_time:28426ms step_avg:161.51ms
step:187/1530 train_loss:4.2338 train_time:28590ms step_avg:161.53ms
step:188/1530 train_loss:4.1702 train_time:28885ms step_avg:162.27ms
step:189/1530 train_loss:4.1059 train_time:29209ms step_avg:163.18ms
step:190/1530 train_loss:4.1935 train_time:29372ms step_avg:163.18ms
step:191/1530 train_loss:4.0716 train_time:29534ms step_avg:163.17ms
step:192/1530 train_loss:4.0258 train_time:29698ms step_avg:163.17ms
step:193/1530 train_loss:4.2489 train_time:29862ms step_avg:163.18ms
step:194/1530 train_loss:4.1691 train_time:30025ms step_avg:163.18ms
step:195/1530 train_loss:4.3511 train_time:30188ms step_avg:163.18ms
step:196/1530 train_loss:4.1796 train_time:30350ms step_avg:163.17ms
step:197/1530 train_loss:4.0370 train_time:30512ms step_avg:163.17ms
step:198/1530 train_loss:4.1766 train_time:30675ms step_avg:163.16ms
step:199/1530 train_loss:4.0351 train_time:30839ms step_avg:163.17ms
step:200/1530 train_loss:4.1080 train_time:31002ms step_avg:163.17ms
step:201/1530 train_loss:4.0135 train_time:31164ms step_avg:163.16ms
step:202/1530 train_loss:4.2597 train_time:31327ms step_avg:163.16ms
step:203/1530 train_loss:4.0655 train_time:31489ms step_avg:163.16ms
step:204/1530 train_loss:4.1902 train_time:31652ms step_avg:163.15ms
step:205/1530 train_loss:4.2406 train_time:31814ms step_avg:163.15ms
step:206/1530 train_loss:3.9479 train_time:31977ms step_avg:163.15ms
step:207/1530 train_loss:4.0864 train_time:32141ms step_avg:163.15ms
step:208/1530 train_loss:4.1000 train_time:32303ms step_avg:163.15ms
step:209/1530 train_loss:4.2384 train_time:32466ms step_avg:163.15ms
step:210/1530 train_loss:4.1747 train_time:32629ms step_avg:163.15ms
step:211/1530 train_loss:4.0612 train_time:32792ms step_avg:163.14ms
step:212/1530 train_loss:4.1259 train_time:32956ms step_avg:163.15ms
step:213/1530 train_loss:4.0523 train_time:33120ms step_avg:163.15ms
step:214/1530 train_loss:4.1236 train_time:33282ms step_avg:163.15ms
step:215/1530 train_loss:3.9809 train_time:33445ms step_avg:163.15ms
step:216/1530 train_loss:4.0027 train_time:33607ms step_avg:163.14ms
step:217/1530 train_loss:4.0111 train_time:33770ms step_avg:163.14ms
step:218/1530 train_loss:4.0884 train_time:33932ms step_avg:163.14ms
step:219/1530 train_loss:4.0805 train_time:34096ms step_avg:163.14ms
step:220/1530 train_loss:4.0805 train_time:34260ms step_avg:163.14ms
step:221/1530 train_loss:4.0905 train_time:34423ms step_avg:163.14ms
step:222/1530 train_loss:3.9911 train_time:34585ms step_avg:163.14ms
step:223/1530 train_loss:3.9949 train_time:34749ms step_avg:163.14ms
step:224/1530 train_loss:4.3047 train_time:34911ms step_avg:163.13ms
step:225/1530 train_loss:3.9339 train_time:35074ms step_avg:163.14ms
step:226/1530 train_loss:3.9892 train_time:35237ms step_avg:163.13ms
step:227/1530 train_loss:3.9762 train_time:35400ms step_avg:163.13ms
step:228/1530 train_loss:4.1465 train_time:35565ms step_avg:163.14ms
step:229/1530 train_loss:3.9286 train_time:35730ms step_avg:163.15ms
step:230/1530 train_loss:4.0360 train_time:35895ms step_avg:163.16ms
step:231/1530 train_loss:3.9031 train_time:36064ms step_avg:163.18ms
step:232/1530 train_loss:3.9767 train_time:36229ms step_avg:163.19ms
step:233/1530 train_loss:4.0925 train_time:36394ms step_avg:163.20ms
step:234/1530 train_loss:4.0274 train_time:36562ms step_avg:163.22ms
step:235/1530 train_loss:3.8989 train_time:36729ms step_avg:163.24ms
step:236/1530 train_loss:4.0801 train_time:36894ms step_avg:163.25ms
step:237/1530 train_loss:4.0769 train_time:37060ms step_avg:163.26ms
step:238/1530 train_loss:3.9423 train_time:37225ms step_avg:163.27ms
step:239/1530 train_loss:4.0775 train_time:37391ms step_avg:163.28ms
step:240/1530 train_loss:4.1131 train_time:37558ms step_avg:163.29ms
step:241/1530 train_loss:3.9611 train_time:37723ms step_avg:163.30ms
step:242/1530 train_loss:4.1399 train_time:37889ms step_avg:163.31ms
step:243/1530 train_loss:4.0074 train_time:38055ms step_avg:163.33ms
step:244/1530 train_loss:4.0782 train_time:38223ms step_avg:163.35ms
step:245/1530 train_loss:4.1402 train_time:38389ms step_avg:163.36ms
step:246/1530 train_loss:4.0608 train_time:38554ms step_avg:163.37ms
step:247/1530 train_loss:4.0028 train_time:38723ms step_avg:163.39ms
step:248/1530 train_loss:4.0931 train_time:38888ms step_avg:163.40ms
step:249/1530 train_loss:3.9191 train_time:39053ms step_avg:163.40ms
step:250/1530 train_loss:3.9809 train_time:39219ms step_avg:163.41ms
step:250/1530 val_loss:4.0102 train_time:39267ms step_avg:163.61ms
step:251/1530 train_loss:4.0834 train_time:39388ms step_avg:163.44ms
step:252/1530 train_loss:4.1614 train_time:39554ms step_avg:163.45ms
step:253/1530 train_loss:3.9375 train_time:39721ms step_avg:163.46ms
step:254/1530 train_loss:3.8776 train_time:39888ms step_avg:163.48ms
step:255/1530 train_loss:4.0747 train_time:40054ms step_avg:163.48ms
step:256/1530 train_loss:3.9951 train_time:40220ms step_avg:163.50ms
step:257/1530 train_loss:3.9874 train_time:40386ms step_avg:163.50ms
step:258/1530 train_loss:3.9878 train_time:40551ms step_avg:163.51ms
step:259/1530 train_loss:4.0364 train_time:40717ms step_avg:163.52ms
step:260/1530 train_loss:4.0661 train_time:40886ms step_avg:163.54ms
step:261/1530 train_loss:4.0300 train_time:41053ms step_avg:163.56ms
step:262/1530 train_loss:3.9974 train_time:41219ms step_avg:163.57ms
step:263/1530 train_loss:3.8970 train_time:41385ms step_avg:163.58ms
step:264/1530 train_loss:3.9855 train_time:41551ms step_avg:163.59ms
step:265/1530 train_loss:3.8699 train_time:41718ms step_avg:163.60ms
step:266/1530 train_loss:3.9175 train_time:41883ms step_avg:163.61ms
step:267/1530 train_loss:3.9257 train_time:42049ms step_avg:163.62ms
step:268/1530 train_loss:3.9573 train_time:42215ms step_avg:163.62ms
step:269/1530 train_loss:3.8553 train_time:42380ms step_avg:163.63ms
step:270/1530 train_loss:4.0960 train_time:42546ms step_avg:163.64ms
step:271/1530 train_loss:3.9683 train_time:42713ms step_avg:163.65ms
step:272/1530 train_loss:3.9323 train_time:42878ms step_avg:163.66ms
step:273/1530 train_loss:3.9458 train_time:43043ms step_avg:163.66ms
step:274/1530 train_loss:4.0381 train_time:43210ms step_avg:163.68ms
step:275/1530 train_loss:4.0583 train_time:43376ms step_avg:163.68ms
step:276/1530 train_loss:4.2298 train_time:43542ms step_avg:163.69ms
step:277/1530 train_loss:4.0372 train_time:43709ms step_avg:163.70ms
step:278/1530 train_loss:4.0920 train_time:43873ms step_avg:163.71ms
step:279/1530 train_loss:4.0050 train_time:44039ms step_avg:163.71ms
step:280/1530 train_loss:4.1826 train_time:44208ms step_avg:163.73ms
step:281/1530 train_loss:3.9704 train_time:44374ms step_avg:163.74ms
step:282/1530 train_loss:3.9368 train_time:44540ms step_avg:163.75ms
step:283/1530 train_loss:3.9085 train_time:44706ms step_avg:163.76ms
step:284/1530 train_loss:4.0426 train_time:44872ms step_avg:163.77ms
step:285/1530 train_loss:4.0613 train_time:45038ms step_avg:163.77ms
step:286/1530 train_loss:4.0900 train_time:45203ms step_avg:163.78ms
step:287/1530 train_loss:3.9184 train_time:45368ms step_avg:163.78ms
step:288/1530 train_loss:4.0175 train_time:45533ms step_avg:163.79ms
step:289/1530 train_loss:3.8646 train_time:45698ms step_avg:163.79ms
step:290/1530 train_loss:3.8625 train_time:45863ms step_avg:163.80ms
step:291/1530 train_loss:3.9090 train_time:46029ms step_avg:163.80ms
step:292/1530 train_loss:3.8691 train_time:46193ms step_avg:163.81ms
step:293/1530 train_loss:3.9071 train_time:46358ms step_avg:163.81ms
step:294/1530 train_loss:3.9422 train_time:46524ms step_avg:163.82ms
step:295/1530 train_loss:3.8393 train_time:46688ms step_avg:163.82ms
step:296/1530 train_loss:3.8597 train_time:46853ms step_avg:163.82ms
step:297/1530 train_loss:3.8660 train_time:47018ms step_avg:163.83ms
step:298/1530 train_loss:3.9691 train_time:47184ms step_avg:163.83ms
step:299/1530 train_loss:3.8266 train_time:47349ms step_avg:163.84ms
step:300/1530 train_loss:3.9718 train_time:47514ms step_avg:163.84ms
step:301/1530 train_loss:3.9647 train_time:47678ms step_avg:163.84ms
step:302/1530 train_loss:3.9374 train_time:47843ms step_avg:163.85ms
step:303/1530 train_loss:3.9838 train_time:48009ms step_avg:163.85ms
step:304/1530 train_loss:3.9695 train_time:48173ms step_avg:163.85ms
step:305/1530 train_loss:4.4543 train_time:48338ms step_avg:163.86ms
step:306/1530 train_loss:3.9365 train_time:48504ms step_avg:163.86ms
step:307/1530 train_loss:3.8366 train_time:48668ms step_avg:163.87ms
step:308/1530 train_loss:3.9820 train_time:48833ms step_avg:163.87ms
step:309/1530 train_loss:3.8642 train_time:48997ms step_avg:163.87ms
step:310/1530 train_loss:4.0891 train_time:49161ms step_avg:163.87ms
step:311/1530 train_loss:3.9238 train_time:49328ms step_avg:163.88ms
step:312/1530 train_loss:3.8615 train_time:49493ms step_avg:163.88ms
step:313/1530 train_loss:3.9430 train_time:49658ms step_avg:163.89ms
step:314/1530 train_loss:4.0654 train_time:49824ms step_avg:163.89ms
step:315/1530 train_loss:3.9431 train_time:49988ms step_avg:163.90ms
step:316/1530 train_loss:3.7950 train_time:50153ms step_avg:163.90ms
step:317/1530 train_loss:3.8732 train_time:50319ms step_avg:163.90ms
step:318/1530 train_loss:3.9190 train_time:50483ms step_avg:163.91ms
step:319/1530 train_loss:3.8828 train_time:50649ms step_avg:163.91ms
step:320/1530 train_loss:4.0169 train_time:50814ms step_avg:163.92ms
step:321/1530 train_loss:3.9631 train_time:50978ms step_avg:163.92ms
step:322/1530 train_loss:3.9383 train_time:51143ms step_avg:163.92ms
step:323/1530 train_loss:4.0124 train_time:51309ms step_avg:163.93ms
step:324/1530 train_loss:3.9493 train_time:51473ms step_avg:163.93ms
step:325/1530 train_loss:4.0269 train_time:51638ms step_avg:163.93ms
step:326/1530 train_loss:3.8902 train_time:51806ms step_avg:163.94ms
step:327/1530 train_loss:4.3940 train_time:51970ms step_avg:163.94ms
step:328/1530 train_loss:4.0702 train_time:52136ms step_avg:163.95ms
step:329/1530 train_loss:3.7934 train_time:52301ms step_avg:163.95ms
step:330/1530 train_loss:3.7543 train_time:52467ms step_avg:163.96ms
step:331/1530 train_loss:3.9733 train_time:52632ms step_avg:163.96ms
step:332/1530 train_loss:3.9121 train_time:52796ms step_avg:163.96ms
step:333/1530 train_loss:3.8895 train_time:52961ms step_avg:163.97ms
step:334/1530 train_loss:3.8393 train_time:53128ms step_avg:163.98ms
step:335/1530 train_loss:4.0138 train_time:53293ms step_avg:163.98ms
step:336/1530 train_loss:3.9527 train_time:53458ms step_avg:163.98ms
step:337/1530 train_loss:4.4280 train_time:53623ms step_avg:163.99ms
step:338/1530 train_loss:3.9429 train_time:53789ms step_avg:163.99ms
step:339/1530 train_loss:3.8712 train_time:53953ms step_avg:163.99ms
step:340/1530 train_loss:3.9297 train_time:54118ms step_avg:164.00ms
step:341/1530 train_loss:3.8585 train_time:54285ms step_avg:164.00ms
step:342/1530 train_loss:3.8156 train_time:54453ms step_avg:164.01ms
step:343/1530 train_loss:3.8380 train_time:54620ms step_avg:164.02ms
step:344/1530 train_loss:3.9944 train_time:54789ms step_avg:164.04ms
step:345/1530 train_loss:3.8194 train_time:54958ms step_avg:164.05ms
step:346/1530 train_loss:3.7704 train_time:55127ms step_avg:164.07ms
step:347/1530 train_loss:3.8022 train_time:55295ms step_avg:164.08ms
step:348/1530 train_loss:3.8601 train_time:55463ms step_avg:164.09ms
step:349/1530 train_loss:3.8315 train_time:55632ms step_avg:164.11ms
step:350/1530 train_loss:3.5683 train_time:55799ms step_avg:164.11ms
step:351/1530 train_loss:3.8266 train_time:55967ms step_avg:164.13ms
step:352/1530 train_loss:4.1802 train_time:56134ms step_avg:164.14ms
step:353/1530 train_loss:3.6566 train_time:56301ms step_avg:164.14ms
step:354/1530 train_loss:3.9294 train_time:56469ms step_avg:164.15ms
step:355/1530 train_loss:3.7894 train_time:56637ms step_avg:164.17ms
step:356/1530 train_loss:3.8861 train_time:56806ms step_avg:164.18ms
step:357/1530 train_loss:3.7558 train_time:56973ms step_avg:164.19ms
step:358/1530 train_loss:3.8600 train_time:57142ms step_avg:164.20ms
step:359/1530 train_loss:3.7919 train_time:57313ms step_avg:164.22ms
step:360/1530 train_loss:3.4346 train_time:57482ms step_avg:164.23ms
step:361/1530 train_loss:4.0236 train_time:57650ms step_avg:164.25ms
step:362/1530 train_loss:3.9197 train_time:57818ms step_avg:164.26ms
step:363/1530 train_loss:3.8412 train_time:57986ms step_avg:164.27ms
step:364/1530 train_loss:3.7469 train_time:58154ms step_avg:164.28ms
step:365/1530 train_loss:3.9169 train_time:58324ms step_avg:164.29ms
step:366/1530 train_loss:3.8626 train_time:58491ms step_avg:164.30ms
step:367/1530 train_loss:3.8627 train_time:58658ms step_avg:164.31ms
step:368/1530 train_loss:3.8562 train_time:58826ms step_avg:164.32ms
step:369/1530 train_loss:3.7490 train_time:58994ms step_avg:164.33ms
step:370/1530 train_loss:3.8784 train_time:59162ms step_avg:164.34ms
step:371/1530 train_loss:3.7351 train_time:59331ms step_avg:164.35ms
step:372/1530 train_loss:3.7022 train_time:59498ms step_avg:164.36ms
step:373/1530 train_loss:3.9077 train_time:59665ms step_avg:164.37ms
step:374/1530 train_loss:3.8299 train_time:59833ms step_avg:164.38ms
step:375/1530 train_loss:3.8068 train_time:60001ms step_avg:164.39ms
step:375/1530 val_loss:3.8271 train_time:60049ms step_avg:164.52ms
step:376/1530 train_loss:3.8641 train_time:60170ms step_avg:164.40ms
step:377/1530 train_loss:3.7884 train_time:60471ms step_avg:164.77ms
step:378/1530 train_loss:3.8442 train_time:60649ms step_avg:164.81ms
step:379/1530 train_loss:3.8755 train_time:60976ms step_avg:165.25ms
step:380/1530 train_loss:3.9581 train_time:61144ms step_avg:165.25ms
step:381/1530 train_loss:3.8405 train_time:61313ms step_avg:165.26ms
step:382/1530 train_loss:3.8007 train_time:61482ms step_avg:165.27ms
step:383/1530 train_loss:3.7949 train_time:61650ms step_avg:165.28ms
step:384/1530 train_loss:3.8738 train_time:61818ms step_avg:165.29ms
step:385/1530 train_loss:3.7967 train_time:61985ms step_avg:165.29ms
step:386/1530 train_loss:3.8909 train_time:62153ms step_avg:165.30ms
step:387/1530 train_loss:4.0490 train_time:62320ms step_avg:165.31ms
step:388/1530 train_loss:3.7938 train_time:62487ms step_avg:165.31ms
step:389/1530 train_loss:3.7988 train_time:62657ms step_avg:165.32ms
step:390/1530 train_loss:3.8957 train_time:62824ms step_avg:165.33ms
step:391/1530 train_loss:3.8149 train_time:62991ms step_avg:165.33ms
step:392/1530 train_loss:3.9192 train_time:63160ms step_avg:165.34ms
step:393/1530 train_loss:3.7633 train_time:63326ms step_avg:165.34ms
step:394/1530 train_loss:3.8855 train_time:63494ms step_avg:165.35ms
step:395/1530 train_loss:3.6292 train_time:63662ms step_avg:165.36ms
step:396/1530 train_loss:3.8336 train_time:63829ms step_avg:165.36ms
step:397/1530 train_loss:3.8730 train_time:63998ms step_avg:165.37ms
step:398/1530 train_loss:3.8892 train_time:64165ms step_avg:165.37ms
step:399/1530 train_loss:3.7678 train_time:64332ms step_avg:165.38ms
step:400/1530 train_loss:3.8290 train_time:64500ms step_avg:165.39ms
step:401/1530 train_loss:3.9200 train_time:64667ms step_avg:165.39ms
step:402/1530 train_loss:3.8463 train_time:64835ms step_avg:165.40ms
step:403/1530 train_loss:3.9606 train_time:65003ms step_avg:165.40ms
step:404/1530 train_loss:3.6784 train_time:65170ms step_avg:165.41ms
step:405/1530 train_loss:3.7895 train_time:65339ms step_avg:165.41ms
step:406/1530 train_loss:4.0881 train_time:65505ms step_avg:165.42ms
step:407/1530 train_loss:3.7802 train_time:65673ms step_avg:165.42ms
step:408/1530 train_loss:3.8214 train_time:65839ms step_avg:165.43ms
step:409/1530 train_loss:3.8524 train_time:66007ms step_avg:165.43ms
step:410/1530 train_loss:3.7550 train_time:66174ms step_avg:165.44ms
step:411/1530 train_loss:3.7585 train_time:66342ms step_avg:165.44ms
step:412/1530 train_loss:4.1781 train_time:66509ms step_avg:165.45ms
step:413/1530 train_loss:3.6259 train_time:66678ms step_avg:165.45ms
step:414/1530 train_loss:4.0116 train_time:66844ms step_avg:165.46ms
step:415/1530 train_loss:3.7574 train_time:67012ms step_avg:165.46ms
step:416/1530 train_loss:3.7678 train_time:67179ms step_avg:165.47ms
step:417/1530 train_loss:3.9552 train_time:67346ms step_avg:165.47ms
step:418/1530 train_loss:3.7009 train_time:67511ms step_avg:165.47ms
step:419/1530 train_loss:3.8085 train_time:67681ms step_avg:165.48ms
step:420/1530 train_loss:3.7064 train_time:67847ms step_avg:165.48ms
step:421/1530 train_loss:3.6516 train_time:68014ms step_avg:165.48ms
step:422/1530 train_loss:3.7775 train_time:68181ms step_avg:165.49ms
step:423/1530 train_loss:3.8733 train_time:68347ms step_avg:165.49ms
step:424/1530 train_loss:3.6134 train_time:68516ms step_avg:165.50ms
step:425/1530 train_loss:3.8000 train_time:68683ms step_avg:165.50ms
step:426/1530 train_loss:3.6531 train_time:68850ms step_avg:165.51ms
step:427/1530 train_loss:3.8870 train_time:69017ms step_avg:165.51ms
step:428/1530 train_loss:3.8112 train_time:69184ms step_avg:165.51ms
step:429/1530 train_loss:3.7538 train_time:69351ms step_avg:165.52ms
step:430/1530 train_loss:3.7063 train_time:69519ms step_avg:165.52ms
step:431/1530 train_loss:3.6299 train_time:69686ms step_avg:165.52ms
step:432/1530 train_loss:3.7676 train_time:69854ms step_avg:165.53ms
step:433/1530 train_loss:3.8212 train_time:70020ms step_avg:165.53ms
step:434/1530 train_loss:3.7762 train_time:70186ms step_avg:165.53ms
step:435/1530 train_loss:3.8058 train_time:70353ms step_avg:165.54ms
step:436/1530 train_loss:3.8318 train_time:70519ms step_avg:165.54ms
step:437/1530 train_loss:3.7218 train_time:70686ms step_avg:165.54ms
step:438/1530 train_loss:3.7012 train_time:70853ms step_avg:165.55ms
step:439/1530 train_loss:3.7166 train_time:71019ms step_avg:165.55ms
step:440/1530 train_loss:3.8926 train_time:71186ms step_avg:165.55ms
step:441/1530 train_loss:3.7623 train_time:71355ms step_avg:165.56ms
step:442/1530 train_loss:3.7390 train_time:71521ms step_avg:165.56ms
step:443/1530 train_loss:3.6219 train_time:71687ms step_avg:165.56ms
step:444/1530 train_loss:3.9277 train_time:71855ms step_avg:165.56ms
step:445/1530 train_loss:3.8412 train_time:72020ms step_avg:165.56ms
step:446/1530 train_loss:3.8304 train_time:72187ms step_avg:165.57ms
step:447/1530 train_loss:3.7555 train_time:72355ms step_avg:165.57ms
step:448/1530 train_loss:3.8484 train_time:72521ms step_avg:165.57ms
step:449/1530 train_loss:3.6923 train_time:72688ms step_avg:165.58ms
step:450/1530 train_loss:3.7292 train_time:72857ms step_avg:165.58ms
step:451/1530 train_loss:3.5906 train_time:73024ms step_avg:165.59ms
step:452/1530 train_loss:3.7076 train_time:73191ms step_avg:165.59ms
step:453/1530 train_loss:3.6727 train_time:73359ms step_avg:165.60ms
step:454/1530 train_loss:3.6384 train_time:73526ms step_avg:165.60ms
step:455/1530 train_loss:3.8399 train_time:73695ms step_avg:165.61ms
step:456/1530 train_loss:3.7247 train_time:73865ms step_avg:165.62ms
step:457/1530 train_loss:3.7822 train_time:74034ms step_avg:165.62ms
step:458/1530 train_loss:3.8282 train_time:74203ms step_avg:165.63ms
step:459/1530 train_loss:3.6327 train_time:74376ms step_avg:165.65ms
step:460/1530 train_loss:3.7971 train_time:74544ms step_avg:165.65ms
step:461/1530 train_loss:3.6946 train_time:74715ms step_avg:165.66ms
step:462/1530 train_loss:3.7364 train_time:74884ms step_avg:165.67ms
step:463/1530 train_loss:3.7780 train_time:75053ms step_avg:165.68ms
step:464/1530 train_loss:3.7122 train_time:75222ms step_avg:165.69ms
step:465/1530 train_loss:3.7193 train_time:75392ms step_avg:165.70ms
step:466/1530 train_loss:3.7945 train_time:75561ms step_avg:165.70ms
step:467/1530 train_loss:3.8240 train_time:75732ms step_avg:165.72ms
step:468/1530 train_loss:3.7883 train_time:75902ms step_avg:165.72ms
step:469/1530 train_loss:3.6845 train_time:76072ms step_avg:165.73ms
step:470/1530 train_loss:3.7636 train_time:76242ms step_avg:165.74ms
step:471/1530 train_loss:3.8097 train_time:76412ms step_avg:165.75ms
step:472/1530 train_loss:3.7836 train_time:76582ms step_avg:165.76ms
step:473/1530 train_loss:3.7133 train_time:76750ms step_avg:165.77ms
step:474/1530 train_loss:3.5928 train_time:76920ms step_avg:165.77ms
step:475/1530 train_loss:4.0216 train_time:77088ms step_avg:165.78ms
step:476/1530 train_loss:3.7563 train_time:77260ms step_avg:165.79ms
step:477/1530 train_loss:3.5968 train_time:77428ms step_avg:165.80ms
step:478/1530 train_loss:3.8227 train_time:77598ms step_avg:165.81ms
step:479/1530 train_loss:3.7732 train_time:77768ms step_avg:165.82ms
step:480/1530 train_loss:3.9270 train_time:77937ms step_avg:165.82ms
step:481/1530 train_loss:3.7231 train_time:78106ms step_avg:165.83ms
step:482/1530 train_loss:3.5303 train_time:78278ms step_avg:165.84ms
step:483/1530 train_loss:3.8041 train_time:78447ms step_avg:165.85ms
step:484/1530 train_loss:3.6648 train_time:78618ms step_avg:165.86ms
step:485/1530 train_loss:3.6584 train_time:78785ms step_avg:165.86ms
step:486/1530 train_loss:3.5713 train_time:78958ms step_avg:165.88ms
step:487/1530 train_loss:3.6825 train_time:79128ms step_avg:165.89ms
step:488/1530 train_loss:3.8755 train_time:79298ms step_avg:165.90ms
step:489/1530 train_loss:3.7186 train_time:79468ms step_avg:165.90ms
step:490/1530 train_loss:3.5958 train_time:79638ms step_avg:165.91ms
step:491/1530 train_loss:3.6185 train_time:79806ms step_avg:165.92ms
step:492/1530 train_loss:3.7324 train_time:79978ms step_avg:165.93ms
step:493/1530 train_loss:3.5737 train_time:80147ms step_avg:165.94ms
step:494/1530 train_loss:3.7039 train_time:80317ms step_avg:165.94ms
step:495/1530 train_loss:3.6616 train_time:80486ms step_avg:165.95ms
step:496/1530 train_loss:3.5203 train_time:80659ms step_avg:165.96ms
step:497/1530 train_loss:3.7365 train_time:80828ms step_avg:165.97ms
step:498/1530 train_loss:3.7914 train_time:80998ms step_avg:165.98ms
step:499/1530 train_loss:3.8193 train_time:81168ms step_avg:165.99ms
step:500/1530 train_loss:3.7353 train_time:81339ms step_avg:166.00ms
step:500/1530 val_loss:3.7040 train_time:81387ms step_avg:166.10ms
step:501/1530 train_loss:3.8047 train_time:81509ms step_avg:166.01ms
step:502/1530 train_loss:3.7444 train_time:81681ms step_avg:166.02ms
step:503/1530 train_loss:3.7773 train_time:81850ms step_avg:166.02ms
step:504/1530 train_loss:3.7199 train_time:82019ms step_avg:166.03ms
step:505/1530 train_loss:3.8059 train_time:82189ms step_avg:166.04ms
step:506/1530 train_loss:3.6375 train_time:82358ms step_avg:166.04ms
step:507/1530 train_loss:3.7595 train_time:82528ms step_avg:166.05ms
step:508/1530 train_loss:3.8212 train_time:82699ms step_avg:166.06ms
step:509/1530 train_loss:3.7662 train_time:82870ms step_avg:166.07ms
step:510/1530 train_loss:3.5819 train_time:83039ms step_avg:166.08ms
step:511/1530 train_loss:3.7775 train_time:83209ms step_avg:166.09ms
step:512/1530 train_loss:3.7154 train_time:83382ms step_avg:166.10ms
step:513/1530 train_loss:3.6718 train_time:83549ms step_avg:166.10ms
step:514/1530 train_loss:3.7892 train_time:83719ms step_avg:166.11ms
step:515/1530 train_loss:3.7395 train_time:83888ms step_avg:166.11ms
step:516/1530 train_loss:4.0690 train_time:84060ms step_avg:166.13ms
step:517/1530 train_loss:3.6904 train_time:84228ms step_avg:166.13ms
step:518/1530 train_loss:3.7703 train_time:84397ms step_avg:166.14ms
step:519/1530 train_loss:3.6522 train_time:84566ms step_avg:166.14ms
step:520/1530 train_loss:3.6847 train_time:84734ms step_avg:166.14ms
step:521/1530 train_loss:3.6633 train_time:84905ms step_avg:166.15ms
step:522/1530 train_loss:3.6605 train_time:85074ms step_avg:166.16ms
step:523/1530 train_loss:4.2930 train_time:85244ms step_avg:166.17ms
step:524/1530 train_loss:3.7371 train_time:85412ms step_avg:166.17ms
step:525/1530 train_loss:3.6784 train_time:85580ms step_avg:166.17ms
step:526/1530 train_loss:3.6943 train_time:85749ms step_avg:166.18ms
step:527/1530 train_loss:3.6568 train_time:85920ms step_avg:166.19ms
step:528/1530 train_loss:3.6333 train_time:86089ms step_avg:166.19ms
step:529/1530 train_loss:3.8440 train_time:86258ms step_avg:166.20ms
step:530/1530 train_loss:3.6430 train_time:86428ms step_avg:166.21ms
step:531/1530 train_loss:3.9191 train_time:86598ms step_avg:166.21ms
step:532/1530 train_loss:3.7373 train_time:86766ms step_avg:166.22ms
step:533/1530 train_loss:3.6544 train_time:86935ms step_avg:166.22ms
step:534/1530 train_loss:3.6696 train_time:87104ms step_avg:166.23ms
step:535/1530 train_loss:3.6070 train_time:87273ms step_avg:166.23ms
step:536/1530 train_loss:3.7468 train_time:87444ms step_avg:166.24ms
step:537/1530 train_loss:3.7195 train_time:87614ms step_avg:166.25ms
step:538/1530 train_loss:3.6217 train_time:87784ms step_avg:166.26ms
step:539/1530 train_loss:4.1116 train_time:87952ms step_avg:166.26ms
step:540/1530 train_loss:3.6787 train_time:88121ms step_avg:166.27ms
step:541/1530 train_loss:3.7836 train_time:88289ms step_avg:166.27ms
step:542/1530 train_loss:3.5859 train_time:88459ms step_avg:166.28ms
step:543/1530 train_loss:3.5809 train_time:88628ms step_avg:166.28ms
step:544/1530 train_loss:3.6345 train_time:88797ms step_avg:166.29ms
step:545/1530 train_loss:3.5932 train_time:88965ms step_avg:166.29ms
step:546/1530 train_loss:3.6257 train_time:89133ms step_avg:166.29ms
step:547/1530 train_loss:3.6399 train_time:89303ms step_avg:166.30ms
step:548/1530 train_loss:3.6089 train_time:89471ms step_avg:166.30ms
step:549/1530 train_loss:3.7223 train_time:89640ms step_avg:166.31ms
step:550/1530 train_loss:3.6166 train_time:89810ms step_avg:166.31ms
step:551/1530 train_loss:3.6258 train_time:89979ms step_avg:166.32ms
step:552/1530 train_loss:3.9382 train_time:90147ms step_avg:166.32ms
step:553/1530 train_loss:3.7592 train_time:90316ms step_avg:166.33ms
step:554/1530 train_loss:3.7116 train_time:90484ms step_avg:166.33ms
step:555/1530 train_loss:3.6280 train_time:90652ms step_avg:166.33ms
step:556/1530 train_loss:3.6960 train_time:90823ms step_avg:166.34ms
step:557/1530 train_loss:3.3109 train_time:90992ms step_avg:166.35ms
step:558/1530 train_loss:3.6117 train_time:91162ms step_avg:166.35ms
step:559/1530 train_loss:3.6478 train_time:91329ms step_avg:166.36ms
step:560/1530 train_loss:3.6890 train_time:91498ms step_avg:166.36ms
step:561/1530 train_loss:3.6149 train_time:91666ms step_avg:166.36ms
step:562/1530 train_loss:3.5589 train_time:91834ms step_avg:166.37ms
step:563/1530 train_loss:3.7579 train_time:92006ms step_avg:166.38ms
step:564/1530 train_loss:3.5752 train_time:92175ms step_avg:166.38ms
step:565/1530 train_loss:3.6797 train_time:92344ms step_avg:166.39ms
step:566/1530 train_loss:3.6232 train_time:92645ms step_avg:166.63ms
step:567/1530 train_loss:3.5989 train_time:92824ms step_avg:166.65ms
step:568/1530 train_loss:3.6840 train_time:92993ms step_avg:166.65ms
step:569/1530 train_loss:3.6506 train_time:93316ms step_avg:166.93ms
step:570/1530 train_loss:3.6856 train_time:93487ms step_avg:166.94ms
step:571/1530 train_loss:3.7560 train_time:93657ms step_avg:166.95ms
step:572/1530 train_loss:3.7246 train_time:93831ms step_avg:166.96ms
step:573/1530 train_loss:3.7365 train_time:94004ms step_avg:166.97ms
step:574/1530 train_loss:3.7716 train_time:94176ms step_avg:166.98ms
step:575/1530 train_loss:3.7326 train_time:94346ms step_avg:166.98ms
step:576/1530 train_loss:3.7623 train_time:94518ms step_avg:166.99ms
step:577/1530 train_loss:3.6709 train_time:94689ms step_avg:167.00ms
step:578/1530 train_loss:3.6743 train_time:94860ms step_avg:167.01ms
step:579/1530 train_loss:3.6684 train_time:95030ms step_avg:167.01ms
step:580/1530 train_loss:3.5869 train_time:95201ms step_avg:167.02ms
step:581/1530 train_loss:3.6395 train_time:95371ms step_avg:167.02ms
step:582/1530 train_loss:3.8520 train_time:95543ms step_avg:167.03ms
step:583/1530 train_loss:3.6249 train_time:95713ms step_avg:167.04ms
step:584/1530 train_loss:3.6018 train_time:95886ms step_avg:167.05ms
step:585/1530 train_loss:3.7900 train_time:96056ms step_avg:167.05ms
step:586/1530 train_loss:3.5178 train_time:96228ms step_avg:167.06ms
step:587/1530 train_loss:3.6717 train_time:96398ms step_avg:167.07ms
step:588/1530 train_loss:3.6419 train_time:96569ms step_avg:167.07ms
step:589/1530 train_loss:3.9961 train_time:96742ms step_avg:167.08ms
step:590/1530 train_loss:3.7807 train_time:96913ms step_avg:167.09ms
step:591/1530 train_loss:3.5053 train_time:97085ms step_avg:167.10ms
step:592/1530 train_loss:3.5376 train_time:97257ms step_avg:167.11ms
step:593/1530 train_loss:3.5010 train_time:97430ms step_avg:167.12ms
step:594/1530 train_loss:3.5556 train_time:97601ms step_avg:167.13ms
step:595/1530 train_loss:3.9112 train_time:97774ms step_avg:167.13ms
step:596/1530 train_loss:3.6474 train_time:97946ms step_avg:167.14ms
step:597/1530 train_loss:3.5843 train_time:98117ms step_avg:167.15ms
step:598/1530 train_loss:3.6596 train_time:98288ms step_avg:167.16ms
step:599/1530 train_loss:3.4762 train_time:98459ms step_avg:167.16ms
step:600/1530 train_loss:3.6016 train_time:98630ms step_avg:167.17ms
step:601/1530 train_loss:3.6468 train_time:98804ms step_avg:167.18ms
step:602/1530 train_loss:3.6708 train_time:98976ms step_avg:167.19ms
step:603/1530 train_loss:3.7809 train_time:99148ms step_avg:167.20ms
step:604/1530 train_loss:3.6092 train_time:99319ms step_avg:167.20ms
step:605/1530 train_loss:3.6116 train_time:99491ms step_avg:167.21ms
step:606/1530 train_loss:3.5718 train_time:99664ms step_avg:167.22ms
step:607/1530 train_loss:3.8354 train_time:99836ms step_avg:167.23ms
step:608/1530 train_loss:3.6330 train_time:100008ms step_avg:167.24ms
step:609/1530 train_loss:3.6180 train_time:100178ms step_avg:167.24ms
step:610/1530 train_loss:3.7035 train_time:100348ms step_avg:167.25ms
step:611/1530 train_loss:3.5991 train_time:100520ms step_avg:167.25ms
step:612/1530 train_loss:3.5729 train_time:100690ms step_avg:167.26ms
step:613/1530 train_loss:3.7656 train_time:100863ms step_avg:167.27ms
step:614/1530 train_loss:3.7003 train_time:101034ms step_avg:167.28ms
step:615/1530 train_loss:3.7019 train_time:101205ms step_avg:167.28ms
step:616/1530 train_loss:3.6285 train_time:101375ms step_avg:167.29ms
step:617/1530 train_loss:3.5551 train_time:101547ms step_avg:167.29ms
step:618/1530 train_loss:3.6844 train_time:101718ms step_avg:167.30ms
step:619/1530 train_loss:3.5499 train_time:101890ms step_avg:167.31ms
step:620/1530 train_loss:3.5908 train_time:102060ms step_avg:167.31ms
step:621/1530 train_loss:3.9200 train_time:102231ms step_avg:167.32ms
step:622/1530 train_loss:3.5716 train_time:102405ms step_avg:167.33ms
step:623/1530 train_loss:3.6039 train_time:102576ms step_avg:167.33ms
step:624/1530 train_loss:3.6899 train_time:102748ms step_avg:167.34ms
step:625/1530 train_loss:3.7024 train_time:102919ms step_avg:167.35ms
step:625/1530 val_loss:3.6226 train_time:102968ms step_avg:167.43ms
step:626/1530 train_loss:3.7343 train_time:103091ms step_avg:167.35ms
step:627/1530 train_loss:3.7159 train_time:103265ms step_avg:167.37ms
step:628/1530 train_loss:3.7639 train_time:103433ms step_avg:167.37ms
step:629/1530 train_loss:3.5947 train_time:103605ms step_avg:167.37ms
step:630/1530 train_loss:3.7242 train_time:103774ms step_avg:167.38ms
step:631/1530 train_loss:3.7408 train_time:103945ms step_avg:167.38ms
step:632/1530 train_loss:3.6482 train_time:104115ms step_avg:167.39ms
step:633/1530 train_loss:3.6020 train_time:104286ms step_avg:167.39ms
step:634/1530 train_loss:3.6997 train_time:104458ms step_avg:167.40ms
step:635/1530 train_loss:3.9503 train_time:104628ms step_avg:167.41ms
step:636/1530 train_loss:3.5491 train_time:104801ms step_avg:167.41ms
step:637/1530 train_loss:3.3576 train_time:104970ms step_avg:167.42ms
step:638/1530 train_loss:3.5917 train_time:105141ms step_avg:167.42ms
step:639/1530 train_loss:3.6335 train_time:105311ms step_avg:167.43ms
step:640/1530 train_loss:3.5701 train_time:105482ms step_avg:167.43ms
step:641/1530 train_loss:3.5894 train_time:105651ms step_avg:167.43ms
step:642/1530 train_loss:3.6331 train_time:105821ms step_avg:167.44ms
step:643/1530 train_loss:3.5989 train_time:105992ms step_avg:167.44ms
step:644/1530 train_loss:3.5562 train_time:106163ms step_avg:167.45ms
step:645/1530 train_loss:3.7707 train_time:106332ms step_avg:167.45ms
step:646/1530 train_loss:3.6751 train_time:106505ms step_avg:167.46ms
step:647/1530 train_loss:3.6621 train_time:106674ms step_avg:167.46ms
step:648/1530 train_loss:3.7127 train_time:106848ms step_avg:167.47ms
step:649/1530 train_loss:3.7663 train_time:107018ms step_avg:167.48ms
step:650/1530 train_loss:3.6251 train_time:107189ms step_avg:167.48ms
step:651/1530 train_loss:3.7696 train_time:107361ms step_avg:167.49ms
step:652/1530 train_loss:3.5825 train_time:107531ms step_avg:167.49ms
step:653/1530 train_loss:3.6609 train_time:107702ms step_avg:167.50ms
step:654/1530 train_loss:3.4255 train_time:107873ms step_avg:167.50ms
step:655/1530 train_loss:3.5772 train_time:108042ms step_avg:167.51ms
step:656/1530 train_loss:3.5750 train_time:108210ms step_avg:167.51ms
step:657/1530 train_loss:3.4956 train_time:108381ms step_avg:167.51ms
step:658/1530 train_loss:3.6869 train_time:108552ms step_avg:167.52ms
step:659/1530 train_loss:3.5884 train_time:108723ms step_avg:167.52ms
step:660/1530 train_loss:3.6818 train_time:108892ms step_avg:167.53ms
step:661/1530 train_loss:3.7546 train_time:109066ms step_avg:167.54ms
step:662/1530 train_loss:3.6740 train_time:109234ms step_avg:167.54ms
step:663/1530 train_loss:3.5529 train_time:109406ms step_avg:167.54ms
step:664/1530 train_loss:3.6009 train_time:109576ms step_avg:167.55ms
step:665/1530 train_loss:3.4923 train_time:109748ms step_avg:167.55ms
step:666/1530 train_loss:3.7769 train_time:109919ms step_avg:167.56ms
step:667/1530 train_loss:3.6037 train_time:110090ms step_avg:167.56ms
step:668/1530 train_loss:3.6517 train_time:110262ms step_avg:167.57ms
step:669/1530 train_loss:3.4862 train_time:110431ms step_avg:167.57ms
step:670/1530 train_loss:3.6007 train_time:110602ms step_avg:167.58ms
step:671/1530 train_loss:3.5570 train_time:110772ms step_avg:167.58ms
step:672/1530 train_loss:3.5732 train_time:110945ms step_avg:167.59ms
step:673/1530 train_loss:3.8474 train_time:111115ms step_avg:167.59ms
step:674/1530 train_loss:3.6242 train_time:111286ms step_avg:167.60ms
step:675/1530 train_loss:3.7114 train_time:111458ms step_avg:167.61ms
step:676/1530 train_loss:3.4912 train_time:111628ms step_avg:167.61ms
step:677/1530 train_loss:3.6021 train_time:111800ms step_avg:167.62ms
step:678/1530 train_loss:3.5539 train_time:111971ms step_avg:167.62ms
step:679/1530 train_loss:3.6826 train_time:112143ms step_avg:167.63ms
step:680/1530 train_loss:3.5884 train_time:112313ms step_avg:167.63ms
step:681/1530 train_loss:3.6142 train_time:112484ms step_avg:167.64ms
step:682/1530 train_loss:3.6549 train_time:112658ms step_avg:167.65ms
step:683/1530 train_loss:3.7396 train_time:112831ms step_avg:167.65ms
step:684/1530 train_loss:3.6518 train_time:113004ms step_avg:167.66ms
step:685/1530 train_loss:3.6840 train_time:113177ms step_avg:167.67ms
step:686/1530 train_loss:3.6415 train_time:113351ms step_avg:167.68ms
step:687/1530 train_loss:3.6696 train_time:113523ms step_avg:167.69ms
step:688/1530 train_loss:3.2146 train_time:113698ms step_avg:167.70ms
step:689/1530 train_loss:3.4136 train_time:113872ms step_avg:167.70ms
step:690/1530 train_loss:3.5397 train_time:114048ms step_avg:167.72ms
step:691/1530 train_loss:3.4117 train_time:114220ms step_avg:167.72ms
step:692/1530 train_loss:3.6338 train_time:114391ms step_avg:167.73ms
step:693/1530 train_loss:3.6505 train_time:114566ms step_avg:167.74ms
step:694/1530 train_loss:3.5536 train_time:114738ms step_avg:167.75ms
step:695/1530 train_loss:3.5337 train_time:114908ms step_avg:167.75ms
step:696/1530 train_loss:3.8551 train_time:115081ms step_avg:167.76ms
step:697/1530 train_loss:3.5856 train_time:115255ms step_avg:167.77ms
step:698/1530 train_loss:3.6381 train_time:115427ms step_avg:167.77ms
step:699/1530 train_loss:3.7778 train_time:115602ms step_avg:167.78ms
step:700/1530 train_loss:3.5747 train_time:115773ms step_avg:167.79ms
step:701/1530 train_loss:3.5482 train_time:115947ms step_avg:167.80ms
step:702/1530 train_loss:3.5134 train_time:116119ms step_avg:167.80ms
step:703/1530 train_loss:3.5024 train_time:116291ms step_avg:167.81ms
step:704/1530 train_loss:3.5768 train_time:116465ms step_avg:167.82ms
step:705/1530 train_loss:3.5615 train_time:116641ms step_avg:167.83ms
step:706/1530 train_loss:3.5825 train_time:116816ms step_avg:167.84ms
step:707/1530 train_loss:3.6506 train_time:116990ms step_avg:167.85ms
step:708/1530 train_loss:3.6072 train_time:117163ms step_avg:167.86ms
step:709/1530 train_loss:3.5823 train_time:117335ms step_avg:167.86ms
step:710/1530 train_loss:3.5359 train_time:117506ms step_avg:167.87ms
step:711/1530 train_loss:3.5882 train_time:117679ms step_avg:167.87ms
step:712/1530 train_loss:3.6476 train_time:117855ms step_avg:167.89ms
step:713/1530 train_loss:3.6580 train_time:118031ms step_avg:167.90ms
step:714/1530 train_loss:3.5606 train_time:118205ms step_avg:167.90ms
step:715/1530 train_loss:3.5742 train_time:118376ms step_avg:167.91ms
step:716/1530 train_loss:3.5892 train_time:118547ms step_avg:167.91ms
step:717/1530 train_loss:3.7073 train_time:118721ms step_avg:167.92ms
step:718/1530 train_loss:3.5955 train_time:118892ms step_avg:167.93ms
step:719/1530 train_loss:3.6771 train_time:119067ms step_avg:167.94ms
step:720/1530 train_loss:3.8318 train_time:119241ms step_avg:167.95ms
step:721/1530 train_loss:3.4658 train_time:119413ms step_avg:167.95ms
step:722/1530 train_loss:3.7371 train_time:119586ms step_avg:167.96ms
step:723/1530 train_loss:3.7696 train_time:119758ms step_avg:167.96ms
step:724/1530 train_loss:3.5701 train_time:119931ms step_avg:167.97ms
step:725/1530 train_loss:3.6504 train_time:120104ms step_avg:167.98ms
step:726/1530 train_loss:3.5329 train_time:120278ms step_avg:167.99ms
step:727/1530 train_loss:3.5835 train_time:120453ms step_avg:168.00ms
step:728/1530 train_loss:3.7280 train_time:120626ms step_avg:168.00ms
step:729/1530 train_loss:3.6705 train_time:120797ms step_avg:168.01ms
step:730/1530 train_loss:3.6628 train_time:120971ms step_avg:168.02ms
step:731/1530 train_loss:3.5500 train_time:121143ms step_avg:168.02ms
step:732/1530 train_loss:3.5950 train_time:121314ms step_avg:168.02ms
step:733/1530 train_loss:3.8298 train_time:121488ms step_avg:168.03ms
step:734/1530 train_loss:3.5569 train_time:121664ms step_avg:168.04ms
step:735/1530 train_loss:3.6194 train_time:121835ms step_avg:168.05ms
step:736/1530 train_loss:3.7338 train_time:122008ms step_avg:168.06ms
step:737/1530 train_loss:3.6759 train_time:122182ms step_avg:168.06ms
step:738/1530 train_loss:3.5965 train_time:122354ms step_avg:168.07ms
step:739/1530 train_loss:3.5053 train_time:122526ms step_avg:168.07ms
step:740/1530 train_loss:4.1156 train_time:122701ms step_avg:168.08ms
step:741/1530 train_loss:3.4933 train_time:122872ms step_avg:168.09ms
step:742/1530 train_loss:3.5555 train_time:123045ms step_avg:168.09ms
step:743/1530 train_loss:3.5838 train_time:123217ms step_avg:168.10ms
step:744/1530 train_loss:3.6444 train_time:123390ms step_avg:168.11ms
step:745/1530 train_loss:3.5890 train_time:123565ms step_avg:168.12ms
step:746/1530 train_loss:3.5920 train_time:123737ms step_avg:168.12ms
step:747/1530 train_loss:3.6441 train_time:123910ms step_avg:168.13ms
step:748/1530 train_loss:3.5653 train_time:124088ms step_avg:168.14ms
step:749/1530 train_loss:3.5663 train_time:124261ms step_avg:168.15ms
step:750/1530 train_loss:3.5983 train_time:124431ms step_avg:168.15ms
step:750/1530 val_loss:3.5651 train_time:124481ms step_avg:168.22ms
step:751/1530 train_loss:3.5690 train_time:124606ms step_avg:168.16ms
step:752/1530 train_loss:3.6163 train_time:124781ms step_avg:168.17ms
step:753/1530 train_loss:3.6172 train_time:124953ms step_avg:168.17ms
step:754/1530 train_loss:3.5941 train_time:125124ms step_avg:168.18ms
step:755/1530 train_loss:3.6790 train_time:125429ms step_avg:168.36ms
step:756/1530 train_loss:3.4570 train_time:125613ms step_avg:168.38ms
step:757/1530 train_loss:3.7224 train_time:125786ms step_avg:168.39ms
step:758/1530 train_loss:3.6477 train_time:125958ms step_avg:168.39ms
step:759/1530 train_loss:3.5918 train_time:126281ms step_avg:168.60ms
step:760/1530 train_loss:3.7089 train_time:126450ms step_avg:168.60ms
step:761/1530 train_loss:3.4037 train_time:126624ms step_avg:168.61ms
step:762/1530 train_loss:3.5458 train_time:126796ms step_avg:168.61ms
step:763/1530 train_loss:3.6658 train_time:126969ms step_avg:168.62ms
step:764/1530 train_loss:3.3183 train_time:127142ms step_avg:168.62ms
step:765/1530 train_loss:3.7295 train_time:127315ms step_avg:168.63ms
step:766/1530 train_loss:3.5727 train_time:127489ms step_avg:168.64ms
step:767/1530 train_loss:3.5659 train_time:127661ms step_avg:168.64ms
step:768/1530 train_loss:3.5742 train_time:127836ms step_avg:168.65ms
step:769/1530 train_loss:3.5908 train_time:128007ms step_avg:168.65ms
step:770/1530 train_loss:3.6431 train_time:128180ms step_avg:168.66ms
step:771/1530 train_loss:3.8800 train_time:128351ms step_avg:168.66ms
step:772/1530 train_loss:3.4514 train_time:128522ms step_avg:168.66ms
step:773/1530 train_loss:3.6302 train_time:128691ms step_avg:168.66ms
step:774/1530 train_loss:3.6425 train_time:128864ms step_avg:168.67ms
step:775/1530 train_loss:3.6049 train_time:129035ms step_avg:168.67ms
step:776/1530 train_loss:3.4023 train_time:129208ms step_avg:168.68ms
step:777/1530 train_loss:3.3917 train_time:129383ms step_avg:168.69ms
step:778/1530 train_loss:3.4908 train_time:129555ms step_avg:168.69ms
step:779/1530 train_loss:3.5782 train_time:129727ms step_avg:168.70ms
step:780/1530 train_loss:3.5911 train_time:129901ms step_avg:168.70ms
step:781/1530 train_loss:3.6733 train_time:130072ms step_avg:168.71ms
step:782/1530 train_loss:3.5846 train_time:130243ms step_avg:168.71ms
step:783/1530 train_loss:3.5663 train_time:130414ms step_avg:168.71ms
step:784/1530 train_loss:3.6101 train_time:130587ms step_avg:168.72ms
step:785/1530 train_loss:3.5612 train_time:130759ms step_avg:168.72ms
step:786/1530 train_loss:3.4373 train_time:130929ms step_avg:168.72ms
step:787/1530 train_loss:3.7515 train_time:131104ms step_avg:168.73ms
step:788/1530 train_loss:3.5016 train_time:131278ms step_avg:168.74ms
step:789/1530 train_loss:3.5505 train_time:131448ms step_avg:168.74ms
step:790/1530 train_loss:3.6280 train_time:131622ms step_avg:168.75ms
step:791/1530 train_loss:3.7675 train_time:131798ms step_avg:168.76ms
step:792/1530 train_loss:3.7580 train_time:131969ms step_avg:168.76ms
step:793/1530 train_loss:3.4422 train_time:132141ms step_avg:168.76ms
step:794/1530 train_loss:3.5964 train_time:132314ms step_avg:168.77ms
step:795/1530 train_loss:3.6750 train_time:132488ms step_avg:168.77ms
step:796/1530 train_loss:3.7367 train_time:132665ms step_avg:168.79ms
step:797/1530 train_loss:3.5268 train_time:132839ms step_avg:168.79ms
step:798/1530 train_loss:3.6419 train_time:133014ms step_avg:168.80ms
step:799/1530 train_loss:3.5379 train_time:133192ms step_avg:168.81ms
step:800/1530 train_loss:3.5275 train_time:133365ms step_avg:168.82ms
step:801/1530 train_loss:3.6287 train_time:133540ms step_avg:168.82ms
step:802/1530 train_loss:3.4929 train_time:133717ms step_avg:168.83ms
step:803/1530 train_loss:3.4870 train_time:133890ms step_avg:168.84ms
step:804/1530 train_loss:3.6213 train_time:134064ms step_avg:168.85ms
step:805/1530 train_loss:3.5142 train_time:134240ms step_avg:168.86ms
step:806/1530 train_loss:3.5609 train_time:134413ms step_avg:168.86ms
step:807/1530 train_loss:3.6417 train_time:134586ms step_avg:168.87ms
step:808/1530 train_loss:3.5433 train_time:134762ms step_avg:168.87ms
step:809/1530 train_loss:3.4948 train_time:134935ms step_avg:168.88ms
step:810/1530 train_loss:3.5593 train_time:135107ms step_avg:168.88ms
step:811/1530 train_loss:3.5821 train_time:135282ms step_avg:168.89ms
step:812/1530 train_loss:3.6007 train_time:135454ms step_avg:168.90ms
step:813/1530 train_loss:3.6252 train_time:135625ms step_avg:168.90ms
step:814/1530 train_loss:3.5692 train_time:135802ms step_avg:168.91ms
step:815/1530 train_loss:3.5609 train_time:135975ms step_avg:168.91ms
step:816/1530 train_loss:3.6801 train_time:136150ms step_avg:168.92ms
step:817/1530 train_loss:3.7680 train_time:136324ms step_avg:168.93ms
step:818/1530 train_loss:3.5259 train_time:136497ms step_avg:168.93ms
step:819/1530 train_loss:3.7174 train_time:136669ms step_avg:168.94ms
step:820/1530 train_loss:3.4920 train_time:136845ms step_avg:168.94ms
step:821/1530 train_loss:3.5572 train_time:137018ms step_avg:168.95ms
step:822/1530 train_loss:3.6905 train_time:137194ms step_avg:168.96ms
step:823/1530 train_loss:3.5730 train_time:137367ms step_avg:168.96ms
step:824/1530 train_loss:3.5106 train_time:137540ms step_avg:168.97ms
step:825/1530 train_loss:3.6164 train_time:137715ms step_avg:168.98ms
step:826/1530 train_loss:3.4744 train_time:137891ms step_avg:168.98ms
step:827/1530 train_loss:3.7366 train_time:138065ms step_avg:168.99ms
step:828/1530 train_loss:3.6195 train_time:138239ms step_avg:169.00ms
step:829/1530 train_loss:3.6282 train_time:138414ms step_avg:169.00ms
step:830/1530 train_loss:3.5356 train_time:138588ms step_avg:169.01ms
step:831/1530 train_loss:3.6000 train_time:138761ms step_avg:169.02ms
step:832/1530 train_loss:3.5139 train_time:138937ms step_avg:169.02ms
step:833/1530 train_loss:3.6497 train_time:139112ms step_avg:169.03ms
step:834/1530 train_loss:3.4715 train_time:139287ms step_avg:169.04ms
step:835/1530 train_loss:3.4565 train_time:139462ms step_avg:169.04ms
step:836/1530 train_loss:3.7170 train_time:139638ms step_avg:169.05ms
step:837/1530 train_loss:3.3971 train_time:139810ms step_avg:169.06ms
step:838/1530 train_loss:3.5947 train_time:139984ms step_avg:169.06ms
step:839/1530 train_loss:3.4199 train_time:140159ms step_avg:169.07ms
step:840/1530 train_loss:3.4665 train_time:140331ms step_avg:169.07ms
step:841/1530 train_loss:3.5682 train_time:140504ms step_avg:169.08ms
step:842/1530 train_loss:3.5827 train_time:140681ms step_avg:169.09ms
step:843/1530 train_loss:3.5586 train_time:140853ms step_avg:169.09ms
step:844/1530 train_loss:3.4344 train_time:141026ms step_avg:169.10ms
step:845/1530 train_loss:3.6639 train_time:141202ms step_avg:169.10ms
step:846/1530 train_loss:3.5126 train_time:141377ms step_avg:169.11ms
step:847/1530 train_loss:3.4917 train_time:141551ms step_avg:169.12ms
step:848/1530 train_loss:3.6405 train_time:141724ms step_avg:169.12ms
step:849/1530 train_loss:3.4903 train_time:141901ms step_avg:169.13ms
step:850/1530 train_loss:3.4432 train_time:142074ms step_avg:169.14ms
step:851/1530 train_loss:3.7328 train_time:142248ms step_avg:169.14ms
step:852/1530 train_loss:3.4385 train_time:142421ms step_avg:169.15ms
step:853/1530 train_loss:3.5691 train_time:142593ms step_avg:169.15ms
step:854/1530 train_loss:3.6483 train_time:142767ms step_avg:169.16ms
step:855/1530 train_loss:3.5186 train_time:142940ms step_avg:169.16ms
step:856/1530 train_loss:3.5462 train_time:143114ms step_avg:169.17ms
step:857/1530 train_loss:3.6034 train_time:143290ms step_avg:169.17ms
step:858/1530 train_loss:3.4687 train_time:143465ms step_avg:169.18ms
step:859/1530 train_loss:3.5583 train_time:143640ms step_avg:169.19ms
step:860/1530 train_loss:3.5864 train_time:143810ms step_avg:169.19ms
step:861/1530 train_loss:3.6328 train_time:143989ms step_avg:169.20ms
step:862/1530 train_loss:3.6026 train_time:144166ms step_avg:169.21ms
step:863/1530 train_loss:3.5729 train_time:144341ms step_avg:169.22ms
step:864/1530 train_loss:3.3785 train_time:144514ms step_avg:169.22ms
step:865/1530 train_loss:3.6029 train_time:144687ms step_avg:169.22ms
step:866/1530 train_loss:3.8891 train_time:144865ms step_avg:169.23ms
step:867/1530 train_loss:3.4587 train_time:145037ms step_avg:169.24ms
step:868/1530 train_loss:3.6433 train_time:145208ms step_avg:169.24ms
step:869/1530 train_loss:3.6131 train_time:145383ms step_avg:169.25ms
step:870/1530 train_loss:3.4447 train_time:145558ms step_avg:169.25ms
step:871/1530 train_loss:3.3869 train_time:145733ms step_avg:169.26ms
step:872/1530 train_loss:3.6485 train_time:145908ms step_avg:169.27ms
step:873/1530 train_loss:3.4614 train_time:146081ms step_avg:169.27ms
step:874/1530 train_loss:3.2177 train_time:146257ms step_avg:169.28ms
step:875/1530 train_loss:3.6346 train_time:146431ms step_avg:169.28ms
step:875/1530 val_loss:3.5184 train_time:146482ms step_avg:169.34ms
step:876/1530 train_loss:3.4346 train_time:146606ms step_avg:169.29ms
step:877/1530 train_loss:3.6233 train_time:146782ms step_avg:169.30ms
step:878/1530 train_loss:3.4722 train_time:146953ms step_avg:169.30ms
step:879/1530 train_loss:3.6539 train_time:147128ms step_avg:169.31ms
step:880/1530 train_loss:3.3091 train_time:147299ms step_avg:169.31ms
step:881/1530 train_loss:3.4806 train_time:147471ms step_avg:169.31ms
step:882/1530 train_loss:3.6953 train_time:147644ms step_avg:169.32ms
step:883/1530 train_loss:3.8415 train_time:147817ms step_avg:169.32ms
step:884/1530 train_loss:3.5697 train_time:147991ms step_avg:169.33ms
step:885/1530 train_loss:3.4875 train_time:148162ms step_avg:169.33ms
step:886/1530 train_loss:3.5746 train_time:148334ms step_avg:169.33ms
step:887/1530 train_loss:4.0792 train_time:148509ms step_avg:169.34ms
step:888/1530 train_loss:3.8319 train_time:148688ms step_avg:169.35ms
step:889/1530 train_loss:3.5215 train_time:148861ms step_avg:169.35ms
step:890/1530 train_loss:3.5389 train_time:149032ms step_avg:169.36ms
step:891/1530 train_loss:3.3608 train_time:149208ms step_avg:169.36ms
step:892/1530 train_loss:3.7131 train_time:149381ms step_avg:169.37ms
step:893/1530 train_loss:3.4207 train_time:149553ms step_avg:169.37ms
step:894/1530 train_loss:3.6335 train_time:149730ms step_avg:169.38ms
step:895/1530 train_loss:3.6775 train_time:149905ms step_avg:169.38ms
step:896/1530 train_loss:3.5004 train_time:150077ms step_avg:169.39ms
step:897/1530 train_loss:3.5479 train_time:150251ms step_avg:169.39ms
step:898/1530 train_loss:3.5901 train_time:150427ms step_avg:169.40ms
step:899/1530 train_loss:3.4766 train_time:150600ms step_avg:169.40ms
step:900/1530 train_loss:3.4243 train_time:150772ms step_avg:169.41ms
step:901/1530 train_loss:3.6168 train_time:150945ms step_avg:169.41ms
step:902/1530 train_loss:3.6333 train_time:151118ms step_avg:169.41ms
step:903/1530 train_loss:3.5403 train_time:151293ms step_avg:169.42ms
step:904/1530 train_loss:3.4961 train_time:151465ms step_avg:169.42ms
step:905/1530 train_loss:3.5016 train_time:151635ms step_avg:169.42ms
step:906/1530 train_loss:3.7025 train_time:151810ms step_avg:169.43ms
step:907/1530 train_loss:3.5172 train_time:151986ms step_avg:169.44ms
step:908/1530 train_loss:3.5638 train_time:152159ms step_avg:169.44ms
step:909/1530 train_loss:3.4578 train_time:152334ms step_avg:169.45ms
step:910/1530 train_loss:3.5242 train_time:152516ms step_avg:169.46ms
step:911/1530 train_loss:3.6427 train_time:152692ms step_avg:169.47ms
step:912/1530 train_loss:3.6010 train_time:152869ms step_avg:169.48ms
step:913/1530 train_loss:3.4610 train_time:153048ms step_avg:169.49ms
step:914/1530 train_loss:3.7414 train_time:153225ms step_avg:169.50ms
step:915/1530 train_loss:3.5360 train_time:153405ms step_avg:169.51ms
step:916/1530 train_loss:3.6160 train_time:153580ms step_avg:169.51ms
step:917/1530 train_loss:3.6006 train_time:153755ms step_avg:169.52ms
step:918/1530 train_loss:4.8302 train_time:153933ms step_avg:169.53ms
step:919/1530 train_loss:3.4978 train_time:154112ms step_avg:169.54ms
step:920/1530 train_loss:3.5858 train_time:154286ms step_avg:169.55ms
step:921/1530 train_loss:3.5477 train_time:154464ms step_avg:169.55ms
step:922/1530 train_loss:3.5819 train_time:154642ms step_avg:169.56ms
step:923/1530 train_loss:3.6103 train_time:154817ms step_avg:169.57ms
step:924/1530 train_loss:3.6819 train_time:154993ms step_avg:169.58ms
step:925/1530 train_loss:3.6494 train_time:155167ms step_avg:169.58ms
step:926/1530 train_loss:3.5534 train_time:155341ms step_avg:169.59ms
step:927/1530 train_loss:3.5526 train_time:155518ms step_avg:169.59ms
step:928/1530 train_loss:3.7774 train_time:155694ms step_avg:169.60ms
step:929/1530 train_loss:3.6126 train_time:155867ms step_avg:169.61ms
step:930/1530 train_loss:3.4033 train_time:156043ms step_avg:169.61ms
step:931/1530 train_loss:3.5008 train_time:156218ms step_avg:169.62ms
step:932/1530 train_loss:3.6491 train_time:156396ms step_avg:169.63ms
step:933/1530 train_loss:3.3678 train_time:156572ms step_avg:169.63ms
step:934/1530 train_loss:3.5782 train_time:156750ms step_avg:169.64ms
step:935/1530 train_loss:3.4378 train_time:156928ms step_avg:169.65ms
step:936/1530 train_loss:3.5169 train_time:157107ms step_avg:169.66ms
step:937/1530 train_loss:3.6212 train_time:157286ms step_avg:169.67ms
step:938/1530 train_loss:3.5412 train_time:157460ms step_avg:169.68ms
step:939/1530 train_loss:3.6720 train_time:157641ms step_avg:169.69ms
step:940/1530 train_loss:3.4816 train_time:157816ms step_avg:169.69ms
step:941/1530 train_loss:3.5483 train_time:157989ms step_avg:169.70ms
step:942/1530 train_loss:3.3548 train_time:158167ms step_avg:169.71ms
step:943/1530 train_loss:3.7120 train_time:158349ms step_avg:169.72ms
step:944/1530 train_loss:3.3964 train_time:158662ms step_avg:169.87ms
step:945/1530 train_loss:3.4284 train_time:158845ms step_avg:169.89ms
step:946/1530 train_loss:5.0674 train_time:159027ms step_avg:169.90ms
step:947/1530 train_loss:3.5952 train_time:159204ms step_avg:169.91ms
step:948/1530 train_loss:3.4837 train_time:159379ms step_avg:169.91ms
step:949/1530 train_loss:3.3755 train_time:159705ms step_avg:170.08ms
step:950/1530 train_loss:3.4421 train_time:159881ms step_avg:170.09ms
step:951/1530 train_loss:3.4060 train_time:160058ms step_avg:170.09ms
step:952/1530 train_loss:3.4744 train_time:160233ms step_avg:170.10ms
step:953/1530 train_loss:3.5719 train_time:160410ms step_avg:170.11ms
step:954/1530 train_loss:3.4449 train_time:160588ms step_avg:170.11ms
step:955/1530 train_loss:3.4753 train_time:160763ms step_avg:170.12ms
step:956/1530 train_loss:3.4461 train_time:160939ms step_avg:170.13ms
step:957/1530 train_loss:3.4948 train_time:161120ms step_avg:170.14ms
step:958/1530 train_loss:3.5059 train_time:161298ms step_avg:170.15ms
step:959/1530 train_loss:3.5122 train_time:161475ms step_avg:170.15ms
step:960/1530 train_loss:3.4083 train_time:161650ms step_avg:170.16ms
step:961/1530 train_loss:3.6433 train_time:161825ms step_avg:170.16ms
step:962/1530 train_loss:3.5935 train_time:162000ms step_avg:170.17ms
step:963/1530 train_loss:3.6568 train_time:162177ms step_avg:170.17ms
step:964/1530 train_loss:3.4323 train_time:162354ms step_avg:170.18ms
step:965/1530 train_loss:3.4779 train_time:162527ms step_avg:170.19ms
step:966/1530 train_loss:3.7108 train_time:162703ms step_avg:170.19ms
step:967/1530 train_loss:3.5217 train_time:162877ms step_avg:170.20ms
step:968/1530 train_loss:3.5128 train_time:163053ms step_avg:170.20ms
step:969/1530 train_loss:3.5839 train_time:163227ms step_avg:170.21ms
step:970/1530 train_loss:3.3773 train_time:163402ms step_avg:170.21ms
step:971/1530 train_loss:3.5322 train_time:163576ms step_avg:170.21ms
step:972/1530 train_loss:3.4713 train_time:163748ms step_avg:170.22ms
step:973/1530 train_loss:3.5412 train_time:163924ms step_avg:170.22ms
step:974/1530 train_loss:3.5902 train_time:164101ms step_avg:170.23ms
step:975/1530 train_loss:3.4660 train_time:164276ms step_avg:170.23ms
step:976/1530 train_loss:3.6692 train_time:164450ms step_avg:170.24ms
step:977/1530 train_loss:3.5728 train_time:164624ms step_avg:170.24ms
step:978/1530 train_loss:3.3603 train_time:164801ms step_avg:170.25ms
step:979/1530 train_loss:3.6258 train_time:164976ms step_avg:170.25ms
step:980/1530 train_loss:3.4142 train_time:165152ms step_avg:170.26ms
step:981/1530 train_loss:3.5762 train_time:165329ms step_avg:170.27ms
step:982/1530 train_loss:3.5374 train_time:165503ms step_avg:170.27ms
step:983/1530 train_loss:3.5150 train_time:165680ms step_avg:170.28ms
step:984/1530 train_loss:3.4988 train_time:165854ms step_avg:170.28ms
step:985/1530 train_loss:3.5764 train_time:166031ms step_avg:170.29ms
step:986/1530 train_loss:3.4198 train_time:166206ms step_avg:170.29ms
step:987/1530 train_loss:3.4846 train_time:166380ms step_avg:170.30ms
step:988/1530 train_loss:3.5131 train_time:166554ms step_avg:170.30ms
step:989/1530 train_loss:3.4203 train_time:166726ms step_avg:170.30ms
step:990/1530 train_loss:3.6605 train_time:166903ms step_avg:170.31ms
step:991/1530 train_loss:3.4665 train_time:167078ms step_avg:170.31ms
step:992/1530 train_loss:3.4452 train_time:167259ms step_avg:170.32ms
step:993/1530 train_loss:3.5057 train_time:167438ms step_avg:170.33ms
step:994/1530 train_loss:3.5963 train_time:167612ms step_avg:170.34ms
step:995/1530 train_loss:3.5303 train_time:167784ms step_avg:170.34ms
step:996/1530 train_loss:3.4584 train_time:167956ms step_avg:170.34ms
step:997/1530 train_loss:3.7549 train_time:168130ms step_avg:170.34ms
step:998/1530 train_loss:3.4419 train_time:168303ms step_avg:170.35ms
step:999/1530 train_loss:3.5840 train_time:168478ms step_avg:170.35ms
step:1000/1530 train_loss:3.4405 train_time:168655ms step_avg:170.36ms
step:1000/1530 val_loss:3.4660 train_time:168706ms step_avg:170.41ms
step:1001/1530 train_loss:3.4962 train_time:168831ms step_avg:170.36ms
step:1002/1530 train_loss:3.3740 train_time:169006ms step_avg:170.37ms
step:1003/1530 train_loss:3.5545 train_time:169182ms step_avg:170.37ms
step:1004/1530 train_loss:3.6005 train_time:169359ms step_avg:170.38ms
step:1005/1530 train_loss:3.3906 train_time:169533ms step_avg:170.39ms
step:1006/1530 train_loss:3.4677 train_time:169712ms step_avg:170.39ms
step:1007/1530 train_loss:3.4381 train_time:169888ms step_avg:170.40ms
step:1008/1530 train_loss:3.5621 train_time:170065ms step_avg:170.41ms
step:1009/1530 train_loss:3.6676 train_time:170244ms step_avg:170.41ms
step:1010/1530 train_loss:3.5587 train_time:170417ms step_avg:170.42ms
step:1011/1530 train_loss:3.5339 train_time:170589ms step_avg:170.42ms
step:1012/1530 train_loss:3.3872 train_time:170763ms step_avg:170.42ms
step:1013/1530 train_loss:3.5357 train_time:170939ms step_avg:170.43ms
step:1014/1530 train_loss:3.6241 train_time:171114ms step_avg:170.43ms
step:1015/1530 train_loss:3.3287 train_time:171290ms step_avg:170.44ms
step:1016/1530 train_loss:3.4101 train_time:171464ms step_avg:170.44ms
step:1017/1530 train_loss:3.4006 train_time:171641ms step_avg:170.45ms
step:1018/1530 train_loss:3.3977 train_time:171816ms step_avg:170.45ms
step:1019/1530 train_loss:3.5220 train_time:171990ms step_avg:170.46ms
step:1020/1530 train_loss:3.3791 train_time:172167ms step_avg:170.46ms
step:1021/1530 train_loss:3.3473 train_time:172342ms step_avg:170.47ms
step:1022/1530 train_loss:3.4792 train_time:172518ms step_avg:170.47ms
step:1023/1530 train_loss:3.5060 train_time:172693ms step_avg:170.48ms
step:1024/1530 train_loss:3.4802 train_time:172869ms step_avg:170.48ms
step:1025/1530 train_loss:3.4775 train_time:173047ms step_avg:170.49ms
step:1026/1530 train_loss:3.6211 train_time:173223ms step_avg:170.50ms
step:1027/1530 train_loss:3.3241 train_time:173400ms step_avg:170.50ms
step:1028/1530 train_loss:3.3934 train_time:173580ms step_avg:170.51ms
step:1029/1530 train_loss:3.3149 train_time:173761ms step_avg:170.52ms
step:1030/1530 train_loss:3.5421 train_time:173936ms step_avg:170.53ms
step:1031/1530 train_loss:3.5050 train_time:174112ms step_avg:170.53ms
step:1032/1530 train_loss:3.6968 train_time:174292ms step_avg:170.54ms
step:1033/1530 train_loss:3.4906 train_time:174468ms step_avg:170.55ms
step:1034/1530 train_loss:3.3929 train_time:174645ms step_avg:170.55ms
step:1035/1530 train_loss:3.4451 train_time:174823ms step_avg:170.56ms
step:1036/1530 train_loss:3.4783 train_time:174998ms step_avg:170.56ms
step:1037/1530 train_loss:3.7930 train_time:175175ms step_avg:170.57ms
step:1038/1530 train_loss:3.6102 train_time:175354ms step_avg:170.58ms
step:1039/1530 train_loss:3.5146 train_time:175537ms step_avg:170.59ms
step:1040/1530 train_loss:3.4122 train_time:175711ms step_avg:170.59ms
step:1041/1530 train_loss:3.4914 train_time:175889ms step_avg:170.60ms
step:1042/1530 train_loss:3.5197 train_time:176063ms step_avg:170.60ms
step:1043/1530 train_loss:3.4475 train_time:176238ms step_avg:170.61ms
step:1044/1530 train_loss:3.4546 train_time:176414ms step_avg:170.61ms
step:1045/1530 train_loss:3.5143 train_time:176590ms step_avg:170.62ms
step:1046/1530 train_loss:3.4261 train_time:176765ms step_avg:170.62ms
step:1047/1530 train_loss:3.6348 train_time:176942ms step_avg:170.63ms
step:1048/1530 train_loss:3.5010 train_time:177119ms step_avg:170.63ms
step:1049/1530 train_loss:3.4031 train_time:177293ms step_avg:170.64ms
step:1050/1530 train_loss:3.3936 train_time:177471ms step_avg:170.65ms
step:1051/1530 train_loss:3.4935 train_time:177648ms step_avg:170.65ms
step:1052/1530 train_loss:3.3643 train_time:177828ms step_avg:170.66ms
step:1053/1530 train_loss:3.6925 train_time:178006ms step_avg:170.67ms
step:1054/1530 train_loss:3.5329 train_time:178184ms step_avg:170.67ms
step:1055/1530 train_loss:3.3847 train_time:178358ms step_avg:170.68ms
step:1056/1530 train_loss:3.4951 train_time:178534ms step_avg:170.68ms
step:1057/1530 train_loss:3.5788 train_time:178711ms step_avg:170.69ms
step:1058/1530 train_loss:3.3054 train_time:178890ms step_avg:170.70ms
step:1059/1530 train_loss:3.3717 train_time:179072ms step_avg:170.71ms
step:1060/1530 train_loss:3.4378 train_time:179248ms step_avg:170.71ms
step:1061/1530 train_loss:3.4188 train_time:179424ms step_avg:170.72ms
step:1062/1530 train_loss:3.3813 train_time:179601ms step_avg:170.72ms
step:1063/1530 train_loss:3.4518 train_time:179775ms step_avg:170.73ms
step:1064/1530 train_loss:3.3806 train_time:179949ms step_avg:170.73ms
step:1065/1530 train_loss:3.3633 train_time:180127ms step_avg:170.74ms
step:1066/1530 train_loss:3.4118 train_time:180303ms step_avg:170.74ms
step:1067/1530 train_loss:3.2780 train_time:180481ms step_avg:170.75ms
step:1068/1530 train_loss:3.4329 train_time:180656ms step_avg:170.75ms
step:1069/1530 train_loss:3.2972 train_time:180838ms step_avg:170.76ms
step:1070/1530 train_loss:3.5666 train_time:181013ms step_avg:170.77ms
step:1071/1530 train_loss:3.5081 train_time:181194ms step_avg:170.78ms
step:1072/1530 train_loss:3.4396 train_time:181368ms step_avg:170.78ms
step:1073/1530 train_loss:3.5202 train_time:181544ms step_avg:170.78ms
step:1074/1530 train_loss:3.4234 train_time:181719ms step_avg:170.79ms
step:1075/1530 train_loss:3.4011 train_time:181897ms step_avg:170.80ms
step:1076/1530 train_loss:3.7959 train_time:182071ms step_avg:170.80ms
step:1077/1530 train_loss:3.4385 train_time:182247ms step_avg:170.80ms
step:1078/1530 train_loss:3.0891 train_time:182431ms step_avg:170.82ms
step:1079/1530 train_loss:3.5301 train_time:182608ms step_avg:170.82ms
step:1080/1530 train_loss:3.4244 train_time:182785ms step_avg:170.83ms
step:1081/1530 train_loss:3.5038 train_time:182959ms step_avg:170.83ms
step:1082/1530 train_loss:3.5917 train_time:183134ms step_avg:170.83ms
step:1083/1530 train_loss:3.4952 train_time:183308ms step_avg:170.84ms
step:1084/1530 train_loss:3.4645 train_time:183483ms step_avg:170.84ms
step:1085/1530 train_loss:3.4346 train_time:183659ms step_avg:170.85ms
step:1086/1530 train_loss:3.6301 train_time:183835ms step_avg:170.85ms
step:1087/1530 train_loss:3.4984 train_time:184009ms step_avg:170.85ms
step:1088/1530 train_loss:3.3662 train_time:184185ms step_avg:170.86ms
step:1089/1530 train_loss:3.3697 train_time:184365ms step_avg:170.87ms
step:1090/1530 train_loss:3.4829 train_time:184544ms step_avg:170.87ms
step:1091/1530 train_loss:3.2882 train_time:184720ms step_avg:170.88ms
step:1092/1530 train_loss:3.4847 train_time:184898ms step_avg:170.89ms
step:1093/1530 train_loss:3.6023 train_time:185072ms step_avg:170.89ms
step:1094/1530 train_loss:3.4509 train_time:185248ms step_avg:170.89ms
step:1095/1530 train_loss:3.4188 train_time:185424ms step_avg:170.90ms
step:1096/1530 train_loss:3.4259 train_time:185601ms step_avg:170.90ms
step:1097/1530 train_loss:3.4906 train_time:185777ms step_avg:170.91ms
step:1098/1530 train_loss:3.5588 train_time:185956ms step_avg:170.92ms
step:1099/1530 train_loss:3.5236 train_time:186135ms step_avg:170.92ms
step:1100/1530 train_loss:3.4265 train_time:186314ms step_avg:170.93ms
step:1101/1530 train_loss:3.2898 train_time:186490ms step_avg:170.93ms
step:1102/1530 train_loss:3.3090 train_time:186669ms step_avg:170.94ms
step:1103/1530 train_loss:3.4442 train_time:186849ms step_avg:170.95ms
step:1104/1530 train_loss:3.3215 train_time:187025ms step_avg:170.95ms
step:1105/1530 train_loss:4.0605 train_time:187203ms step_avg:170.96ms
step:1106/1530 train_loss:3.2223 train_time:187379ms step_avg:170.97ms
step:1107/1530 train_loss:3.5665 train_time:187554ms step_avg:170.97ms
step:1108/1530 train_loss:3.3482 train_time:187727ms step_avg:170.97ms
step:1109/1530 train_loss:3.5071 train_time:187902ms step_avg:170.98ms
step:1110/1530 train_loss:3.4265 train_time:188075ms step_avg:170.98ms
step:1111/1530 train_loss:3.4874 train_time:188251ms step_avg:170.98ms
step:1112/1530 train_loss:3.5586 train_time:188432ms step_avg:170.99ms
step:1113/1530 train_loss:3.4262 train_time:188616ms step_avg:171.00ms
step:1114/1530 train_loss:3.3687 train_time:188794ms step_avg:171.01ms
step:1115/1530 train_loss:3.2399 train_time:188973ms step_avg:171.02ms
step:1116/1530 train_loss:3.4277 train_time:189147ms step_avg:171.02ms
step:1117/1530 train_loss:3.5873 train_time:189326ms step_avg:171.03ms
step:1118/1530 train_loss:3.6195 train_time:189503ms step_avg:171.03ms
step:1119/1530 train_loss:3.4796 train_time:189676ms step_avg:171.03ms
step:1120/1530 train_loss:3.4878 train_time:189852ms step_avg:171.04ms
step:1121/1530 train_loss:3.3883 train_time:190030ms step_avg:171.04ms
step:1122/1530 train_loss:3.4564 train_time:190204ms step_avg:171.05ms
step:1123/1530 train_loss:3.5769 train_time:190381ms step_avg:171.05ms
step:1124/1530 train_loss:3.3378 train_time:190555ms step_avg:171.06ms
step:1125/1530 train_loss:3.2344 train_time:190732ms step_avg:171.06ms
step:1125/1530 val_loss:3.4081 train_time:190782ms step_avg:171.10ms
step:1126/1530 train_loss:3.4768 train_time:190910ms step_avg:171.07ms
step:1127/1530 train_loss:3.6725 train_time:191089ms step_avg:171.07ms
step:1128/1530 train_loss:3.2296 train_time:191265ms step_avg:171.08ms
step:1129/1530 train_loss:3.5581 train_time:191445ms step_avg:171.09ms
step:1130/1530 train_loss:3.3770 train_time:191622ms step_avg:171.09ms
step:1131/1530 train_loss:3.3952 train_time:191805ms step_avg:171.10ms
step:1132/1530 train_loss:3.3657 train_time:191978ms step_avg:171.10ms
step:1133/1530 train_loss:3.4905 train_time:192287ms step_avg:171.23ms
step:1134/1530 train_loss:3.4489 train_time:192470ms step_avg:171.24ms
step:1135/1530 train_loss:3.5186 train_time:192649ms step_avg:171.24ms
step:1136/1530 train_loss:3.5607 train_time:192827ms step_avg:171.25ms
step:1137/1530 train_loss:3.4580 train_time:193003ms step_avg:171.25ms
step:1138/1530 train_loss:3.3572 train_time:193183ms step_avg:171.26ms
step:1139/1530 train_loss:3.6554 train_time:193513ms step_avg:171.40ms
step:1140/1530 train_loss:3.4556 train_time:193690ms step_avg:171.41ms
step:1141/1530 train_loss:3.5928 train_time:193868ms step_avg:171.41ms
step:1142/1530 train_loss:3.4452 train_time:194045ms step_avg:171.42ms
step:1143/1530 train_loss:3.3639 train_time:194223ms step_avg:171.42ms
step:1144/1530 train_loss:3.4388 train_time:194399ms step_avg:171.43ms
step:1145/1530 train_loss:3.5897 train_time:194574ms step_avg:171.43ms
step:1146/1530 train_loss:3.5543 train_time:194755ms step_avg:171.44ms
step:1147/1530 train_loss:3.4850 train_time:194933ms step_avg:171.44ms
step:1148/1530 train_loss:3.4969 train_time:195110ms step_avg:171.45ms
step:1149/1530 train_loss:3.3211 train_time:195291ms step_avg:171.46ms
step:1150/1530 train_loss:3.3713 train_time:195467ms step_avg:171.46ms
step:1151/1530 train_loss:3.3191 train_time:195647ms step_avg:171.47ms
step:1152/1530 train_loss:3.3949 train_time:195827ms step_avg:171.48ms
step:1153/1530 train_loss:3.4346 train_time:196006ms step_avg:171.48ms
step:1154/1530 train_loss:3.5216 train_time:196180ms step_avg:171.49ms
step:1155/1530 train_loss:3.3178 train_time:196361ms step_avg:171.49ms
step:1156/1530 train_loss:3.5344 train_time:196543ms step_avg:171.50ms
step:1157/1530 train_loss:3.4944 train_time:196720ms step_avg:171.51ms
step:1158/1530 train_loss:3.2478 train_time:196896ms step_avg:171.51ms
step:1159/1530 train_loss:3.3507 train_time:197073ms step_avg:171.52ms
step:1160/1530 train_loss:3.3374 train_time:197246ms step_avg:171.52ms
step:1161/1530 train_loss:3.0757 train_time:197424ms step_avg:171.52ms
step:1162/1530 train_loss:3.4240 train_time:197600ms step_avg:171.53ms
step:1163/1530 train_loss:3.3946 train_time:197779ms step_avg:171.53ms
step:1164/1530 train_loss:3.2930 train_time:197957ms step_avg:171.54ms
step:1165/1530 train_loss:3.2449 train_time:198132ms step_avg:171.54ms
step:1166/1530 train_loss:3.3818 train_time:198311ms step_avg:171.55ms
step:1167/1530 train_loss:3.4110 train_time:198486ms step_avg:171.55ms
step:1168/1530 train_loss:3.7249 train_time:198662ms step_avg:171.56ms
step:1169/1530 train_loss:3.3765 train_time:198839ms step_avg:171.56ms
step:1170/1530 train_loss:3.3930 train_time:199017ms step_avg:171.57ms
step:1171/1530 train_loss:3.3104 train_time:199193ms step_avg:171.57ms
step:1172/1530 train_loss:3.4217 train_time:199367ms step_avg:171.57ms
step:1173/1530 train_loss:3.5382 train_time:199548ms step_avg:171.58ms
step:1174/1530 train_loss:3.3817 train_time:199734ms step_avg:171.59ms
step:1175/1530 train_loss:3.3593 train_time:199913ms step_avg:171.60ms
step:1176/1530 train_loss:3.4213 train_time:200093ms step_avg:171.61ms
step:1177/1530 train_loss:3.4458 train_time:200274ms step_avg:171.61ms
step:1178/1530 train_loss:3.4983 train_time:200449ms step_avg:171.62ms
step:1179/1530 train_loss:3.3976 train_time:200625ms step_avg:171.62ms
step:1180/1530 train_loss:3.3533 train_time:200810ms step_avg:171.63ms
step:1181/1530 train_loss:3.3382 train_time:200988ms step_avg:171.64ms
step:1182/1530 train_loss:3.3743 train_time:201166ms step_avg:171.64ms
step:1183/1530 train_loss:3.3318 train_time:201344ms step_avg:171.65ms
step:1184/1530 train_loss:3.5073 train_time:201521ms step_avg:171.65ms
step:1185/1530 train_loss:3.5388 train_time:201702ms step_avg:171.66ms
step:1186/1530 train_loss:3.3648 train_time:201882ms step_avg:171.67ms
step:1187/1530 train_loss:3.4159 train_time:202067ms step_avg:171.68ms
step:1188/1530 train_loss:3.4435 train_time:202243ms step_avg:171.68ms
step:1189/1530 train_loss:3.2774 train_time:202422ms step_avg:171.69ms
step:1190/1530 train_loss:3.4405 train_time:202600ms step_avg:171.70ms
step:1191/1530 train_loss:3.5838 train_time:202780ms step_avg:171.70ms
step:1192/1530 train_loss:3.3928 train_time:202958ms step_avg:171.71ms
step:1193/1530 train_loss:3.2739 train_time:203134ms step_avg:171.71ms
step:1194/1530 train_loss:3.5580 train_time:203312ms step_avg:171.72ms
step:1195/1530 train_loss:3.3698 train_time:203492ms step_avg:171.72ms
step:1196/1530 train_loss:3.3862 train_time:203678ms step_avg:171.74ms
step:1197/1530 train_loss:3.2899 train_time:203858ms step_avg:171.74ms
step:1198/1530 train_loss:3.2964 train_time:204042ms step_avg:171.75ms
step:1199/1530 train_loss:3.3432 train_time:204221ms step_avg:171.76ms
step:1200/1530 train_loss:3.4511 train_time:204397ms step_avg:171.76ms
step:1201/1530 train_loss:3.4774 train_time:204575ms step_avg:171.77ms
step:1202/1530 train_loss:3.6273 train_time:204765ms step_avg:171.78ms
step:1203/1530 train_loss:3.4025 train_time:204944ms step_avg:171.79ms
step:1204/1530 train_loss:3.3024 train_time:205124ms step_avg:171.80ms
step:1205/1530 train_loss:3.4376 train_time:205300ms step_avg:171.80ms
step:1206/1530 train_loss:3.4732 train_time:205477ms step_avg:171.80ms
step:1207/1530 train_loss:3.5149 train_time:205657ms step_avg:171.81ms
step:1208/1530 train_loss:3.3935 train_time:205832ms step_avg:171.81ms
step:1209/1530 train_loss:3.2458 train_time:206010ms step_avg:171.82ms
step:1210/1530 train_loss:3.3006 train_time:206189ms step_avg:171.82ms
step:1211/1530 train_loss:3.3945 train_time:206368ms step_avg:171.83ms
step:1212/1530 train_loss:3.3940 train_time:206545ms step_avg:171.83ms
step:1213/1530 train_loss:3.4108 train_time:206724ms step_avg:171.84ms
step:1214/1530 train_loss:3.2548 train_time:206905ms step_avg:171.85ms
step:1215/1530 train_loss:3.3950 train_time:207080ms step_avg:171.85ms
step:1216/1530 train_loss:3.3330 train_time:207259ms step_avg:171.86ms
step:1217/1530 train_loss:3.3171 train_time:207436ms step_avg:171.86ms
step:1218/1530 train_loss:3.4101 train_time:207612ms step_avg:171.86ms
step:1219/1530 train_loss:3.2567 train_time:207795ms step_avg:171.87ms
step:1220/1530 train_loss:3.4734 train_time:207970ms step_avg:171.88ms
step:1221/1530 train_loss:3.5050 train_time:208145ms step_avg:171.88ms
step:1222/1530 train_loss:3.4276 train_time:208320ms step_avg:171.88ms
step:1223/1530 train_loss:3.2913 train_time:208498ms step_avg:171.89ms
step:1224/1530 train_loss:3.2527 train_time:208680ms step_avg:171.89ms
step:1225/1530 train_loss:3.3661 train_time:208858ms step_avg:171.90ms
step:1226/1530 train_loss:3.3332 train_time:209038ms step_avg:171.91ms
step:1227/1530 train_loss:3.2787 train_time:209218ms step_avg:171.91ms
step:1228/1530 train_loss:3.4475 train_time:209394ms step_avg:171.92ms
step:1229/1530 train_loss:3.3692 train_time:209573ms step_avg:171.92ms
step:1230/1530 train_loss:3.4014 train_time:209757ms step_avg:171.93ms
step:1231/1530 train_loss:3.5781 train_time:209937ms step_avg:171.94ms
step:1232/1530 train_loss:3.4990 train_time:210117ms step_avg:171.95ms
step:1233/1530 train_loss:3.4282 train_time:210295ms step_avg:171.95ms
step:1234/1530 train_loss:3.5878 train_time:210473ms step_avg:171.96ms
step:1235/1530 train_loss:3.3213 train_time:210655ms step_avg:171.96ms
step:1236/1530 train_loss:3.2920 train_time:210830ms step_avg:171.97ms
step:1237/1530 train_loss:3.2748 train_time:211007ms step_avg:171.97ms
step:1238/1530 train_loss:3.2750 train_time:211191ms step_avg:171.98ms
step:1239/1530 train_loss:3.3318 train_time:211368ms step_avg:171.98ms
step:1240/1530 train_loss:3.3839 train_time:211548ms step_avg:171.99ms
step:1241/1530 train_loss:3.4247 train_time:211727ms step_avg:172.00ms
step:1242/1530 train_loss:3.2994 train_time:211905ms step_avg:172.00ms
step:1243/1530 train_loss:3.3991 train_time:212084ms step_avg:172.01ms
step:1244/1530 train_loss:3.4015 train_time:212257ms step_avg:172.01ms
step:1245/1530 train_loss:3.4125 train_time:212433ms step_avg:172.01ms
step:1246/1530 train_loss:3.2445 train_time:212610ms step_avg:172.01ms
step:1247/1530 train_loss:3.3762 train_time:212786ms step_avg:172.02ms
step:1248/1530 train_loss:3.4252 train_time:212963ms step_avg:172.02ms
step:1249/1530 train_loss:3.4256 train_time:213142ms step_avg:172.03ms
step:1250/1530 train_loss:3.3017 train_time:213321ms step_avg:172.03ms
step:1250/1530 val_loss:3.3544 train_time:213376ms step_avg:172.08ms
step:1251/1530 train_loss:3.4962 train_time:213508ms step_avg:172.05ms
step:1252/1530 train_loss:3.3590 train_time:213685ms step_avg:172.05ms
step:1253/1530 train_loss:3.3123 train_time:213862ms step_avg:172.05ms
step:1254/1530 train_loss:3.4184 train_time:214043ms step_avg:172.06ms
step:1255/1530 train_loss:3.5135 train_time:214231ms step_avg:172.07ms
step:1256/1530 train_loss:3.3043 train_time:214413ms step_avg:172.08ms
step:1257/1530 train_loss:3.3723 train_time:214591ms step_avg:172.09ms
step:1258/1530 train_loss:3.3686 train_time:214775ms step_avg:172.10ms
step:1259/1530 train_loss:3.3307 train_time:214955ms step_avg:172.10ms
step:1260/1530 train_loss:3.2070 train_time:215130ms step_avg:172.10ms
step:1261/1530 train_loss:3.3027 train_time:215308ms step_avg:172.11ms
step:1262/1530 train_loss:3.3239 train_time:215491ms step_avg:172.12ms
step:1263/1530 train_loss:3.2353 train_time:215671ms step_avg:172.12ms
step:1264/1530 train_loss:3.4430 train_time:215848ms step_avg:172.13ms
step:1265/1530 train_loss:3.4250 train_time:216023ms step_avg:172.13ms
step:1266/1530 train_loss:3.4380 train_time:216203ms step_avg:172.14ms
step:1267/1530 train_loss:3.3689 train_time:216383ms step_avg:172.14ms
step:1268/1530 train_loss:3.4087 train_time:216562ms step_avg:172.15ms
step:1269/1530 train_loss:3.2531 train_time:216747ms step_avg:172.16ms
step:1270/1530 train_loss:3.1065 train_time:216924ms step_avg:172.16ms
step:1271/1530 train_loss:3.4042 train_time:217102ms step_avg:172.17ms
step:1272/1530 train_loss:3.3507 train_time:217278ms step_avg:172.17ms
step:1273/1530 train_loss:3.3768 train_time:217461ms step_avg:172.18ms
step:1274/1530 train_loss:3.3605 train_time:217642ms step_avg:172.19ms
step:1275/1530 train_loss:3.4361 train_time:217819ms step_avg:172.19ms
step:1276/1530 train_loss:3.4709 train_time:217994ms step_avg:172.19ms
step:1277/1530 train_loss:3.4132 train_time:218173ms step_avg:172.20ms
step:1278/1530 train_loss:3.4071 train_time:218349ms step_avg:172.20ms
step:1279/1530 train_loss:3.2650 train_time:218530ms step_avg:172.21ms
step:1280/1530 train_loss:3.3605 train_time:218714ms step_avg:172.22ms
step:1281/1530 train_loss:3.4256 train_time:218891ms step_avg:172.22ms
step:1282/1530 train_loss:3.4690 train_time:219064ms step_avg:172.22ms
step:1283/1530 train_loss:3.3346 train_time:219245ms step_avg:172.23ms
step:1284/1530 train_loss:3.3666 train_time:219424ms step_avg:172.23ms
step:1285/1530 train_loss:3.3638 train_time:219601ms step_avg:172.24ms
step:1286/1530 train_loss:3.3361 train_time:219778ms step_avg:172.24ms
step:1287/1530 train_loss:3.4906 train_time:219957ms step_avg:172.24ms
step:1288/1530 train_loss:3.2936 train_time:220137ms step_avg:172.25ms
step:1289/1530 train_loss:3.3830 train_time:220323ms step_avg:172.26ms
step:1290/1530 train_loss:3.4586 train_time:220507ms step_avg:172.27ms
step:1291/1530 train_loss:3.3864 train_time:220687ms step_avg:172.28ms
step:1292/1530 train_loss:3.4797 train_time:220869ms step_avg:172.28ms
step:1293/1530 train_loss:3.5160 train_time:221051ms step_avg:172.29ms
step:1294/1530 train_loss:3.4602 train_time:221231ms step_avg:172.30ms
step:1295/1530 train_loss:3.2876 train_time:221411ms step_avg:172.30ms
step:1296/1530 train_loss:3.3738 train_time:221592ms step_avg:172.31ms
step:1297/1530 train_loss:3.2772 train_time:221772ms step_avg:172.32ms
step:1298/1530 train_loss:3.2717 train_time:221954ms step_avg:172.32ms
step:1299/1530 train_loss:3.3987 train_time:222133ms step_avg:172.33ms
step:1300/1530 train_loss:3.4031 train_time:222309ms step_avg:172.33ms
step:1301/1530 train_loss:3.4046 train_time:222486ms step_avg:172.34ms
step:1302/1530 train_loss:3.5720 train_time:222670ms step_avg:172.34ms
step:1303/1530 train_loss:3.3064 train_time:222851ms step_avg:172.35ms
step:1304/1530 train_loss:3.5153 train_time:223034ms step_avg:172.36ms
step:1305/1530 train_loss:3.2560 train_time:223210ms step_avg:172.36ms
step:1306/1530 train_loss:3.4497 train_time:223391ms step_avg:172.37ms
step:1307/1530 train_loss:3.4572 train_time:223566ms step_avg:172.37ms
step:1308/1530 train_loss:3.2857 train_time:223745ms step_avg:172.38ms
step:1309/1530 train_loss:3.3114 train_time:223924ms step_avg:172.38ms
step:1310/1530 train_loss:3.2879 train_time:224101ms step_avg:172.39ms
step:1311/1530 train_loss:3.2970 train_time:224279ms step_avg:172.39ms
step:1312/1530 train_loss:3.3732 train_time:224460ms step_avg:172.40ms
step:1313/1530 train_loss:3.3468 train_time:224637ms step_avg:172.40ms
step:1314/1530 train_loss:3.0435 train_time:224818ms step_avg:172.41ms
step:1315/1530 train_loss:3.2808 train_time:224994ms step_avg:172.41ms
step:1316/1530 train_loss:3.4018 train_time:225171ms step_avg:172.41ms
step:1317/1530 train_loss:3.4174 train_time:225349ms step_avg:172.42ms
step:1318/1530 train_loss:3.3049 train_time:225535ms step_avg:172.43ms
step:1319/1530 train_loss:3.4282 train_time:225715ms step_avg:172.43ms
step:1320/1530 train_loss:3.4641 train_time:225897ms step_avg:172.44ms
step:1321/1530 train_loss:3.3692 train_time:226076ms step_avg:172.45ms
step:1322/1530 train_loss:3.3252 train_time:226386ms step_avg:172.55ms
step:1323/1530 train_loss:3.3195 train_time:226575ms step_avg:172.56ms
step:1324/1530 train_loss:3.4338 train_time:226753ms step_avg:172.57ms
step:1325/1530 train_loss:3.4935 train_time:226937ms step_avg:172.58ms
step:1326/1530 train_loss:3.2121 train_time:227118ms step_avg:172.58ms
step:1327/1530 train_loss:3.1685 train_time:227293ms step_avg:172.58ms
step:1328/1530 train_loss:3.4944 train_time:227473ms step_avg:172.59ms
step:1329/1530 train_loss:3.2980 train_time:227815ms step_avg:172.72ms
step:1330/1530 train_loss:3.4293 train_time:227994ms step_avg:172.72ms
step:1331/1530 train_loss:3.3280 train_time:228170ms step_avg:172.73ms
step:1332/1530 train_loss:3.7434 train_time:228352ms step_avg:172.73ms
step:1333/1530 train_loss:3.4791 train_time:228534ms step_avg:172.74ms
step:1334/1530 train_loss:3.3755 train_time:228711ms step_avg:172.74ms
step:1335/1530 train_loss:3.2935 train_time:228889ms step_avg:172.75ms
step:1336/1530 train_loss:3.2973 train_time:229074ms step_avg:172.76ms
step:1337/1530 train_loss:3.5538 train_time:229253ms step_avg:172.76ms
step:1338/1530 train_loss:3.5245 train_time:229433ms step_avg:172.77ms
step:1339/1530 train_loss:3.3358 train_time:229612ms step_avg:172.77ms
step:1340/1530 train_loss:3.2862 train_time:229790ms step_avg:172.77ms
step:1341/1530 train_loss:3.5955 train_time:229965ms step_avg:172.78ms
step:1342/1530 train_loss:3.3586 train_time:230146ms step_avg:172.78ms
step:1343/1530 train_loss:3.3569 train_time:230324ms step_avg:172.79ms
step:1344/1530 train_loss:3.4177 train_time:230503ms step_avg:172.79ms
step:1345/1530 train_loss:3.3828 train_time:230685ms step_avg:172.80ms
step:1346/1530 train_loss:3.2993 train_time:230860ms step_avg:172.80ms
step:1347/1530 train_loss:3.2792 train_time:231039ms step_avg:172.80ms
step:1348/1530 train_loss:3.3518 train_time:231218ms step_avg:172.81ms
step:1349/1530 train_loss:3.2757 train_time:231393ms step_avg:172.81ms
step:1350/1530 train_loss:3.3926 train_time:231573ms step_avg:172.82ms
step:1351/1530 train_loss:3.2487 train_time:231749ms step_avg:172.82ms
step:1352/1530 train_loss:3.3091 train_time:231926ms step_avg:172.82ms
step:1353/1530 train_loss:3.4051 train_time:232106ms step_avg:172.83ms
step:1354/1530 train_loss:3.2608 train_time:232284ms step_avg:172.83ms
step:1355/1530 train_loss:3.1872 train_time:232461ms step_avg:172.83ms
step:1356/1530 train_loss:3.5113 train_time:232641ms step_avg:172.84ms
step:1357/1530 train_loss:3.4250 train_time:232821ms step_avg:172.84ms
step:1358/1530 train_loss:3.1877 train_time:233000ms step_avg:172.85ms
step:1359/1530 train_loss:3.4425 train_time:233179ms step_avg:172.85ms
step:1360/1530 train_loss:3.3540 train_time:233358ms step_avg:172.86ms
step:1361/1530 train_loss:3.1349 train_time:233542ms step_avg:172.87ms
step:1362/1530 train_loss:3.3982 train_time:233723ms step_avg:172.87ms
step:1363/1530 train_loss:3.2822 train_time:233911ms step_avg:172.88ms
step:1364/1530 train_loss:3.2976 train_time:234087ms step_avg:172.89ms
step:1365/1530 train_loss:3.3137 train_time:234266ms step_avg:172.89ms
step:1366/1530 train_loss:3.4250 train_time:234447ms step_avg:172.90ms
step:1367/1530 train_loss:3.3953 train_time:234626ms step_avg:172.90ms
step:1368/1530 train_loss:3.3460 train_time:234805ms step_avg:172.90ms
step:1369/1530 train_loss:3.2788 train_time:234993ms step_avg:172.92ms
step:1370/1530 train_loss:3.6067 train_time:235176ms step_avg:172.92ms
step:1371/1530 train_loss:3.3155 train_time:235358ms step_avg:172.93ms
step:1372/1530 train_loss:3.3695 train_time:235541ms step_avg:172.94ms
step:1373/1530 train_loss:3.3716 train_time:235721ms step_avg:172.94ms
step:1374/1530 train_loss:3.1495 train_time:235901ms step_avg:172.95ms
step:1375/1530 train_loss:3.5322 train_time:236081ms step_avg:172.95ms
step:1375/1530 val_loss:3.3129 train_time:236131ms step_avg:172.99ms
step:1376/1530 train_loss:3.3502 train_time:236259ms step_avg:172.96ms
step:1377/1530 train_loss:3.4856 train_time:236439ms step_avg:172.96ms
step:1378/1530 train_loss:3.4734 train_time:236616ms step_avg:172.96ms
step:1379/1530 train_loss:3.1195 train_time:236797ms step_avg:172.97ms
step:1380/1530 train_loss:3.3148 train_time:236976ms step_avg:172.97ms
step:1381/1530 train_loss:3.7073 train_time:237161ms step_avg:172.98ms
step:1382/1530 train_loss:3.2131 train_time:237340ms step_avg:172.99ms
step:1383/1530 train_loss:3.3923 train_time:237522ms step_avg:173.00ms
step:1384/1530 train_loss:3.4782 train_time:237705ms step_avg:173.00ms
step:1385/1530 train_loss:3.4055 train_time:237879ms step_avg:173.00ms
step:1386/1530 train_loss:3.3406 train_time:238060ms step_avg:173.01ms
step:1387/1530 train_loss:3.1999 train_time:238240ms step_avg:173.01ms
step:1388/1530 train_loss:3.3477 train_time:238417ms step_avg:173.02ms
step:1389/1530 train_loss:3.3184 train_time:238599ms step_avg:173.02ms
step:1390/1530 train_loss:3.5675 train_time:238775ms step_avg:173.03ms
step:1391/1530 train_loss:3.2918 train_time:238953ms step_avg:173.03ms
step:1392/1530 train_loss:3.2908 train_time:239133ms step_avg:173.03ms
step:1393/1530 train_loss:3.2407 train_time:239313ms step_avg:173.04ms
step:1394/1530 train_loss:3.4985 train_time:239490ms step_avg:173.04ms
step:1395/1530 train_loss:3.3942 train_time:239668ms step_avg:173.05ms
step:1396/1530 train_loss:3.4089 train_time:239846ms step_avg:173.05ms
step:1397/1530 train_loss:3.3134 train_time:240021ms step_avg:173.05ms
step:1398/1530 train_loss:3.2581 train_time:240196ms step_avg:173.05ms
step:1399/1530 train_loss:3.3169 train_time:240376ms step_avg:173.06ms
step:1400/1530 train_loss:3.3194 train_time:240560ms step_avg:173.06ms
step:1401/1530 train_loss:3.3467 train_time:240737ms step_avg:173.07ms
step:1402/1530 train_loss:3.3002 train_time:240916ms step_avg:173.07ms
step:1403/1530 train_loss:3.5003 train_time:241099ms step_avg:173.08ms
step:1404/1530 train_loss:3.2814 train_time:241277ms step_avg:173.08ms
step:1405/1530 train_loss:3.3159 train_time:241457ms step_avg:173.09ms
step:1406/1530 train_loss:3.3130 train_time:241640ms step_avg:173.09ms
step:1407/1530 train_loss:3.1758 train_time:241816ms step_avg:173.10ms
step:1408/1530 train_loss:3.3149 train_time:241995ms step_avg:173.10ms
step:1409/1530 train_loss:3.3021 train_time:242184ms step_avg:173.11ms
step:1410/1530 train_loss:3.2890 train_time:242361ms step_avg:173.12ms
step:1411/1530 train_loss:3.3629 train_time:242538ms step_avg:173.12ms
step:1412/1530 train_loss:3.3401 train_time:242715ms step_avg:173.12ms
step:1413/1530 train_loss:3.3659 train_time:242894ms step_avg:173.12ms
step:1414/1530 train_loss:3.3314 train_time:243075ms step_avg:173.13ms
step:1415/1530 train_loss:3.4106 train_time:243261ms step_avg:173.14ms
step:1416/1530 train_loss:3.2325 train_time:243451ms step_avg:173.15ms
step:1417/1530 train_loss:3.2812 train_time:243633ms step_avg:173.16ms
step:1418/1530 train_loss:3.3934 train_time:243813ms step_avg:173.16ms
step:1419/1530 train_loss:3.3509 train_time:243996ms step_avg:173.17ms
step:1420/1530 train_loss:3.3701 train_time:244177ms step_avg:173.18ms
step:1421/1530 train_loss:3.3742 train_time:244355ms step_avg:173.18ms
step:1422/1530 train_loss:3.3342 train_time:244533ms step_avg:173.18ms
step:1423/1530 train_loss:3.3157 train_time:244712ms step_avg:173.19ms
step:1424/1530 train_loss:3.3369 train_time:244895ms step_avg:173.19ms
step:1425/1530 train_loss:3.1929 train_time:245081ms step_avg:173.20ms
step:1426/1530 train_loss:3.3270 train_time:245259ms step_avg:173.21ms
step:1427/1530 train_loss:3.2872 train_time:245442ms step_avg:173.21ms
step:1428/1530 train_loss:3.3785 train_time:245619ms step_avg:173.22ms
step:1429/1530 train_loss:3.3544 train_time:245797ms step_avg:173.22ms
step:1430/1530 train_loss:3.2634 train_time:245978ms step_avg:173.22ms
step:1431/1530 train_loss:3.3188 train_time:246160ms step_avg:173.23ms
step:1432/1530 train_loss:3.3404 train_time:246341ms step_avg:173.24ms
step:1433/1530 train_loss:3.1342 train_time:246525ms step_avg:173.24ms
step:1434/1530 train_loss:3.2872 train_time:246708ms step_avg:173.25ms
step:1435/1530 train_loss:3.1180 train_time:246888ms step_avg:173.25ms
step:1436/1530 train_loss:3.2324 train_time:247067ms step_avg:173.26ms
step:1437/1530 train_loss:3.4116 train_time:247244ms step_avg:173.26ms
step:1438/1530 train_loss:3.3865 train_time:247419ms step_avg:173.26ms
step:1439/1530 train_loss:3.3167 train_time:247599ms step_avg:173.27ms
step:1440/1530 train_loss:3.1955 train_time:247775ms step_avg:173.27ms
step:1441/1530 train_loss:3.3413 train_time:247955ms step_avg:173.27ms
step:1442/1530 train_loss:3.3942 train_time:248137ms step_avg:173.28ms
step:1443/1530 train_loss:3.4931 train_time:248325ms step_avg:173.29ms
step:1444/1530 train_loss:3.4497 train_time:248502ms step_avg:173.29ms
step:1445/1530 train_loss:3.3388 train_time:248680ms step_avg:173.30ms
step:1446/1530 train_loss:3.2022 train_time:248859ms step_avg:173.30ms
step:1447/1530 train_loss:3.2977 train_time:249039ms step_avg:173.31ms
step:1448/1530 train_loss:3.2973 train_time:249217ms step_avg:173.31ms
step:1449/1530 train_loss:3.3976 train_time:249395ms step_avg:173.31ms
step:1450/1530 train_loss:3.3904 train_time:249575ms step_avg:173.32ms
step:1451/1530 train_loss:3.2075 train_time:249754ms step_avg:173.32ms
step:1452/1530 train_loss:3.3308 train_time:249934ms step_avg:173.32ms
step:1453/1530 train_loss:3.2633 train_time:250109ms step_avg:173.33ms
step:1454/1530 train_loss:3.2936 train_time:250287ms step_avg:173.33ms
step:1455/1530 train_loss:3.3316 train_time:250470ms step_avg:173.34ms
step:1456/1530 train_loss:3.2872 train_time:250648ms step_avg:173.34ms
step:1457/1530 train_loss:3.1581 train_time:250823ms step_avg:173.34ms
step:1458/1530 train_loss:3.4263 train_time:251000ms step_avg:173.34ms
step:1459/1530 train_loss:3.2710 train_time:251181ms step_avg:173.35ms
step:1460/1530 train_loss:3.3160 train_time:251360ms step_avg:173.35ms
step:1461/1530 train_loss:3.4304 train_time:251539ms step_avg:173.36ms
step:1462/1530 train_loss:3.2624 train_time:251716ms step_avg:173.36ms
step:1463/1530 train_loss:3.4662 train_time:251898ms step_avg:173.36ms
step:1464/1530 train_loss:3.3601 train_time:252076ms step_avg:173.37ms
step:1465/1530 train_loss:3.3589 train_time:252256ms step_avg:173.37ms
step:1466/1530 train_loss:3.2892 train_time:252434ms step_avg:173.37ms
step:1467/1530 train_loss:3.3959 train_time:252614ms step_avg:173.38ms
step:1468/1530 train_loss:3.2890 train_time:252791ms step_avg:173.38ms
step:1469/1530 train_loss:3.2772 train_time:252971ms step_avg:173.39ms
step:1470/1530 train_loss:3.3313 train_time:253153ms step_avg:173.39ms
step:1471/1530 train_loss:3.2616 train_time:253339ms step_avg:173.40ms
step:1472/1530 train_loss:3.2549 train_time:253523ms step_avg:173.41ms
step:1473/1530 train_loss:3.4467 train_time:253700ms step_avg:173.41ms
step:1474/1530 train_loss:3.3135 train_time:253884ms step_avg:173.42ms
step:1475/1530 train_loss:3.1526 train_time:254069ms step_avg:173.43ms
step:1476/1530 train_loss:3.2645 train_time:254248ms step_avg:173.43ms
step:1477/1530 train_loss:3.2372 train_time:254433ms step_avg:173.44ms
step:1478/1530 train_loss:3.3123 train_time:254616ms step_avg:173.44ms
step:1479/1530 train_loss:3.4014 train_time:254797ms step_avg:173.45ms
step:1480/1530 train_loss:3.2723 train_time:254976ms step_avg:173.45ms
step:1481/1530 train_loss:3.4514 train_time:255158ms step_avg:173.46ms
step:1482/1530 train_loss:3.3668 train_time:255345ms step_avg:173.47ms
step:1483/1530 train_loss:3.2790 train_time:255536ms step_avg:173.48ms
step:1484/1530 train_loss:3.2682 train_time:255724ms step_avg:173.49ms
step:1485/1530 train_loss:3.2848 train_time:255904ms step_avg:173.49ms
step:1486/1530 train_loss:3.2314 train_time:256090ms step_avg:173.50ms
step:1487/1530 train_loss:3.3404 train_time:256272ms step_avg:173.51ms
step:1488/1530 train_loss:3.2447 train_time:256455ms step_avg:173.52ms
step:1489/1530 train_loss:3.3159 train_time:256636ms step_avg:173.52ms
step:1490/1530 train_loss:3.2553 train_time:256816ms step_avg:173.52ms
step:1491/1530 train_loss:3.1612 train_time:256997ms step_avg:173.53ms
step:1492/1530 train_loss:3.2688 train_time:257177ms step_avg:173.53ms
step:1493/1530 train_loss:3.4345 train_time:257356ms step_avg:173.54ms
step:1494/1530 train_loss:3.2998 train_time:257534ms step_avg:173.54ms
step:1495/1530 train_loss:3.0303 train_time:257718ms step_avg:173.55ms
step:1496/1530 train_loss:3.3629 train_time:257901ms step_avg:173.55ms
step:1497/1530 train_loss:3.3114 train_time:258088ms step_avg:173.56ms
step:1498/1530 train_loss:3.3479 train_time:258272ms step_avg:173.57ms
step:1499/1530 train_loss:3.3149 train_time:258461ms step_avg:173.58ms
step:1500/1530 train_loss:3.2925 train_time:258655ms step_avg:173.59ms
step:1500/1530 val_loss:3.2814 train_time:258711ms step_avg:173.63ms
step:1501/1530 train_loss:3.0899 train_time:258848ms step_avg:173.61ms
step:1502/1530 train_loss:3.3616 train_time:259040ms step_avg:173.62ms
step:1503/1530 train_loss:3.2411 train_time:259219ms step_avg:173.62ms
step:1504/1530 train_loss:3.2521 train_time:259400ms step_avg:173.63ms
step:1505/1530 train_loss:3.2114 train_time:259578ms step_avg:173.63ms
step:1506/1530 train_loss:3.2814 train_time:259760ms step_avg:173.64ms
step:1507/1530 train_loss:3.1809 train_time:259957ms step_avg:173.65ms
step:1508/1530 train_loss:3.4867 train_time:260141ms step_avg:173.66ms
step:1509/1530 train_loss:3.2820 train_time:260319ms step_avg:173.66ms
step:1510/1530 train_loss:3.2751 train_time:260499ms step_avg:173.67ms
step:1511/1530 train_loss:3.4184 train_time:260815ms step_avg:173.76ms
step:1512/1530 train_loss:3.4194 train_time:261002ms step_avg:173.77ms
step:1513/1530 train_loss:3.2676 train_time:261184ms step_avg:173.78ms
step:1514/1530 train_loss:3.0863 train_time:261364ms step_avg:173.78ms
step:1515/1530 train_loss:3.2440 train_time:261545ms step_avg:173.78ms
step:1516/1530 train_loss:3.2568 train_time:261730ms step_avg:173.79ms
step:1517/1530 train_loss:3.2997 train_time:261910ms step_avg:173.80ms
step:1518/1530 train_loss:3.2047 train_time:262094ms step_avg:173.80ms
step:1519/1530 train_loss:3.5122 train_time:262426ms step_avg:173.91ms
step:1520/1530 train_loss:3.1319 train_time:262609ms step_avg:173.91ms
step:1521/1530 train_loss:3.2048 train_time:262786ms step_avg:173.92ms
step:1522/1530 train_loss:3.3570 train_time:262968ms step_avg:173.92ms
step:1523/1530 train_loss:3.2319 train_time:263146ms step_avg:173.92ms
step:1524/1530 train_loss:3.3514 train_time:263328ms step_avg:173.93ms
step:1525/1530 train_loss:3.3365 train_time:263515ms step_avg:173.94ms
step:1526/1530 train_loss:3.2812 train_time:263705ms step_avg:173.95ms
step:1527/1530 train_loss:3.2906 train_time:263886ms step_avg:173.95ms
step:1528/1530 train_loss:3.4087 train_time:264065ms step_avg:173.96ms
step:1529/1530 train_loss:3.4108 train_time:264244ms step_avg:173.96ms
step:1530/1530 train_loss:3.2389 train_time:264422ms step_avg:173.96ms
step:1530/1530 val_loss:3.2790 train_time:264476ms step_avg:174.00ms