Add test for deepseek_math #1148
Open
+86
−0
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
meenakshiramanathan1
requested review from
vkovinicTT and
mstojkovicTT
as code owners
meenakshiramanathan1
requested review from
nvukobratTT,
pilkicTT and
dgolubovicTT
|
||||||||||||||
|
||||||||||||||
meenakshiramanathan1
force-pushed
the
mramanathan/deepseek_math
branch
from
4adc46d to
676ee1f
Compare
|
||||||||||||||
1 similar comment
|
||||||||||||||
|
||||||||||||||
1 similar comment
|
||||||||||||||
nvukobratTT
reviewed
Feb 3, 2025
forge/test/models/pytorch/multimodal/deepseek/test_deepseek_math.py
Outdated
Show resolved
Hide resolved
nvukobratTT
reviewed
Feb 3, 2025
forge/test/models/pytorch/multimodal/deepseek/test_deepseek_math.py
Outdated
Show resolved
Hide resolved
meenakshiramanathan1
force-pushed
the
mramanathan/deepseek_math
branch
2 times, most recently
from
1a22034 to
4af6e34
Compare
|
||||||||||||||
1 similar comment
|
||||||||||||||
|
||||||||||||||
1 similar comment
|
||||||||||||||
vkovinicTT
requested changes
Feb 4, 2025
forge/test/models/pytorch/multimodal/deepseek/test_deepseek_math.py
Outdated
Show resolved
Hide resolved
meenakshiramanathan1
force-pushed
the
mramanathan/deepseek_math
branch
from
4af6e34 to
573b5a3
Compare
|
||||||||||||||
|
||||||||||||||
|
||||||||||||||
|
||||||||||||||
vkovinicTT
approved these changes
Feb 4, 2025
Comment on lines
+13
to
+43
| def generation(max_new_tokens, compiled_model, input_ids, tokenizer): | ||
| for i in range(max_new_tokens): | ||
| logits = compiled_model(input_ids) | ||
| next_token_logits = logits[:, -1, :] | ||
| next_token_id = torch.argmax(next_token_logits, dim=-1) | ||
|
|
||
| if next_token_id == tokenizer.eos_token_id: | ||
| break | ||
|
|
||
| input_ids = torch.cat([input_ids, next_token_id.unsqueeze(0)], dim=-1) | ||
|
|
||
| generated_text = tokenizer.decode(input_ids[0], skip_special_tokens=True) | ||
| return generated_text | ||
|
|
||
|
|
||
| def download_model_and_tokenizer(model_name): | ||
| tokenizer = AutoTokenizer.from_pretrained(model_name) | ||
| model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, device_map="cpu") | ||
| model.generation_config = GenerationConfig.from_pretrained(model_name) | ||
| model.generation_config.pad_token_id = model.generation_config.eos_token_id | ||
| model.generation_config.use_cache = False | ||
|
|
||
| # Prepare input sentence | ||
| messages = [ | ||
| { | ||
| "role": "user", | ||
| "content": "what is the integral of x^2 from 0 to 2?\nPlease reason step by step, and put your final answer within \\boxed{}.", | ||
| } | ||
| ] | ||
| input_ids = tokenizer.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt") | ||
| return model, tokenizer, input_ids |
There was a problem hiding this comment.
Can we just move this to utils folder.
mstojkovicTT
reviewed
Feb 4, 2025
| next_token_logits = logits[:, -1, :] | ||
| next_token_id = torch.argmax(next_token_logits, dim=-1) | ||
|
|
||
| if next_token_id == tokenizer.eos_token_id: |
There was a problem hiding this comment.
Comparing a tensor to an integer can work when the tensor has a single element, but it’s clearer and safer to extract the scalar value :))
Something like this should work:
next_token_id.item() == tokenizer.eos_token_id:
| return generated_text | ||
|
|
||
|
|
||
| def download_model_and_tokenizer(model_name): |
There was a problem hiding this comment.
One suggestion—would it be possible to add an option to set use_cache to True? We might consider an approach similar to what’s shown here. Also, as @vkovinicTT mentioned, this function might fit better in the utils folder :))
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.