From 09fcd151550ed81482c911722bfc31049a0e541e Mon Sep 17 00:00:00 2001 From: Yuta Nagano <52748151+yutanagano@users.noreply.github.com> Date: Wed, 8 May 2024 16:08:10 +0100 Subject: [PATCH 1/2] rename olga and unpaired variants to synthetic data and shuffled data respectively --- .../{SCEPTR_unpaired => SCEPTR_shuffled_data}/config.json | 2 +- .../{SCEPTR_unpaired => SCEPTR_shuffled_data}/log.csv | 0 .../state_dict.pt | 0 .../{SCEPTR_OLGA => SCEPTR_synthetic_data}/config.json | 2 +- .../{SCEPTR_OLGA => SCEPTR_synthetic_data}/log.csv | 0 .../{SCEPTR_OLGA => SCEPTR_synthetic_data}/state_dict.pt | 0 src/sceptr/variant.py | 8 ++++---- tests/test_variants.py | 4 ++-- 8 files changed, 8 insertions(+), 8 deletions(-) rename src/sceptr/_model_saves/{SCEPTR_unpaired => SCEPTR_shuffled_data}/config.json (98%) rename src/sceptr/_model_saves/{SCEPTR_unpaired => SCEPTR_shuffled_data}/log.csv (100%) rename src/sceptr/_model_saves/{SCEPTR_unpaired => SCEPTR_shuffled_data}/state_dict.pt (100%) rename src/sceptr/_model_saves/{SCEPTR_OLGA => SCEPTR_synthetic_data}/config.json (98%) rename src/sceptr/_model_saves/{SCEPTR_OLGA => SCEPTR_synthetic_data}/log.csv (100%) rename src/sceptr/_model_saves/{SCEPTR_OLGA => SCEPTR_synthetic_data}/state_dict.pt (100%) diff --git a/src/sceptr/_model_saves/SCEPTR_unpaired/config.json b/src/sceptr/_model_saves/SCEPTR_shuffled_data/config.json similarity index 98% rename from src/sceptr/_model_saves/SCEPTR_unpaired/config.json rename to src/sceptr/_model_saves/SCEPTR_shuffled_data/config.json index 0c226d8..d748d0f 100644 --- a/src/sceptr/_model_saves/SCEPTR_unpaired/config.json +++ b/src/sceptr/_model_saves/SCEPTR_shuffled_data/config.json @@ -4,7 +4,7 @@ "initargs": {} }, "model": { - "name": "SCEPTR (unpaired)", + "name": "SCEPTR (shuffled data)", "path_to_pretrained_state_dict": null, "token_embedder": { "class": "CdrSimpleEmbedder", diff --git a/src/sceptr/_model_saves/SCEPTR_unpaired/log.csv b/src/sceptr/_model_saves/SCEPTR_shuffled_data/log.csv similarity index 100% rename from src/sceptr/_model_saves/SCEPTR_unpaired/log.csv rename to src/sceptr/_model_saves/SCEPTR_shuffled_data/log.csv diff --git a/src/sceptr/_model_saves/SCEPTR_unpaired/state_dict.pt b/src/sceptr/_model_saves/SCEPTR_shuffled_data/state_dict.pt similarity index 100% rename from src/sceptr/_model_saves/SCEPTR_unpaired/state_dict.pt rename to src/sceptr/_model_saves/SCEPTR_shuffled_data/state_dict.pt diff --git a/src/sceptr/_model_saves/SCEPTR_OLGA/config.json b/src/sceptr/_model_saves/SCEPTR_synthetic_data/config.json similarity index 98% rename from src/sceptr/_model_saves/SCEPTR_OLGA/config.json rename to src/sceptr/_model_saves/SCEPTR_synthetic_data/config.json index 645952c..6bdb9e9 100644 --- a/src/sceptr/_model_saves/SCEPTR_OLGA/config.json +++ b/src/sceptr/_model_saves/SCEPTR_synthetic_data/config.json @@ -4,7 +4,7 @@ "initargs": {} }, "model": { - "name": "SCEPTR (OLGA)", + "name": "SCEPTR (synthetic data)", "path_to_pretrained_state_dict": null, "token_embedder": { "class": "CdrSimpleEmbedder", diff --git a/src/sceptr/_model_saves/SCEPTR_OLGA/log.csv b/src/sceptr/_model_saves/SCEPTR_synthetic_data/log.csv similarity index 100% rename from src/sceptr/_model_saves/SCEPTR_OLGA/log.csv rename to src/sceptr/_model_saves/SCEPTR_synthetic_data/log.csv diff --git a/src/sceptr/_model_saves/SCEPTR_OLGA/state_dict.pt b/src/sceptr/_model_saves/SCEPTR_synthetic_data/state_dict.pt similarity index 100% rename from src/sceptr/_model_saves/SCEPTR_OLGA/state_dict.pt rename to src/sceptr/_model_saves/SCEPTR_synthetic_data/state_dict.pt diff --git a/src/sceptr/variant.py b/src/sceptr/variant.py index ef37a19..61ffdc1 100644 --- a/src/sceptr/variant.py +++ b/src/sceptr/variant.py @@ -45,12 +45,12 @@ def average_pooling(): return load_variant("SCEPTR_average_pooling") -def unpaired(): - return load_variant("SCEPTR_unpaired") +def shuffled_data(): + return load_variant("SCEPTR_shuffled_data") -def olga(): - return load_variant("SCEPTR_OLGA") +def synthetic_data(): + return load_variant("SCEPTR_synthetic_data") def dropout_noise_only(): diff --git a/tests/test_variants.py b/tests/test_variants.py index 8de890e..b60c6bf 100644 --- a/tests/test_variants.py +++ b/tests/test_variants.py @@ -25,8 +25,8 @@ def dummy_data(): variant.tiny(), variant.blosum(), variant.average_pooling(), - variant.unpaired(), - variant.olga(), + variant.shuffled_data(), + variant.synthetic_data(), variant.dropout_noise_only(), variant.finetuned(), variant.a_sceptr(), From 8ea38fadb12ee95cca81fbb292ce96780df55b56 Mon Sep 17 00:00:00 2001 From: Yuta Nagano <52748151+yutanagano@users.noreply.github.com> Date: Wed, 8 May 2024 16:09:37 +0100 Subject: [PATCH 2/2] Update README to reflect changes in variant naming --- README.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index 8b9091c..698b884 100644 --- a/README.md +++ b/README.md @@ -160,8 +160,8 @@ This `Sceptr` object will then have the methods: `calc_pdist_vector`, `calc_cdis |`sceptr.variant.tiny`|smaller variant with model dimensionality 16| |`sceptr.variant.blosum`|variant using BLOSUM62 embeddings instead of one-hot| |`sceptr.variant.average_pooling`|variant using the average-pooling method to generate the TCR representation vector| -|`sceptr.variant.unpaired`|variant trained on the Tanno et al. dataset with randomised alpha/beta pairing| -|`sceptr.variant.olga`|variant trained using synthetic TCR sequences generated by OLGA| +|`sceptr.variant.shuffled_data`|variant trained on the Tanno et al. dataset with randomised alpha/beta pairing| +|`sceptr.variant.synthetic_data`|variant trained using synthetic TCR sequences generated by OLGA| |`sceptr.variant.dropout_noise_only`|variant trained without residue/chain dropping during autocontrastive learning| |`sceptr.variant.finetuned`|variant fine-tuned using supervised contrastive learning for six pMHCs with peptides GILGFVFTL, NLVPMVATV, SPRWYFYYL, TFEYVSQPFLMDLE, TTDPSFLGRY and YLQPRTFLL (from [VDJdb](https://vdjdb.cdr3.net/))|