Upload PrototypeContrastiveForMaskedLM

config.json

@@ -0,0 +1,105 @@
+{
+  "_expbin_vocab_size": 1003,
+  "_expnum_vocab_size": 1000,
+  "_name_or_path": "/home/xinyu402/projects/rrg-bengioy-ad/xinyu402/scZero_ckpts/240202_181731/checkpoint-40000",
+  "_tf_class_vocab_size": 39,
+  "_tf_superclass_vocab_size": 15,
+  "_vocab_size": 25427,
+  "additional_force_clsprotcontrast_rankbased": null,
+  "architectures": [
+    "PrototypeContrastiveForMaskedLM"
+  ],
+  "attention_probs_dropout_prob": 0.02,
+  "bert_objective": "plm",
+  "bugfix_binary_ce_tau": "learnable",
+  "bugfix_rank_contrastive_diagonal": null,
+  "bugfix_rank_contrastive_snnloss_tau": null,
+  "cell_level_tau": 0.01,
+  "cell_ontology_type": "co_geometry",
+  "centroid_init": null,
+  "classifier_dropout": null,
+  "clm_as_mlm": null,
+  "cls_prot_affine_projection": 1,
+  "cls_prot_contrast_change_tau": null,
+  "contrastive_dropout": null,
+  "ct_alpha": 0.9,
+  "ctgraph_thresh": 0.0001,
+  "expbin_tieweight": 0,
+  "expbin_vocab_size": null,
+  "expnum_vocab_size": null,
+  "fixbug_rank_contrastive": 1,
+  "force_centroid_norm": null,
+  "force_clsprotcontrast_rankbased": null,
+  "global_contrast": 0,
+  "global_contrast_only_for_cell_wise": 0,
+  "global_contrast_only_for_sample_wise": 0,
+  "gnn_type": "gcn",
+  "gradient_checkpointing": false,
+  "hidden_act": "relu",
+  "hidden_dropout_prob": 0.02,
+  "hidden_size": 256,
+  "initialize_decoder_as_identity": null,
+  "initializer_range": 0.02,
+  "intermediate_size": 512,
+  "is_cell_wise": 1,
+  "is_cell_wise_ctgraph_sim": 1,
+  "is_cell_wise_forbidden_ancestor": 1,
+  "is_cls_prot_contrast_toallcelltype": 1,
+  "is_newdata": "frac100",
+  "is_prototype_wise": 1,
+  "is_prototype_wise_allcelltypetoallcelltype_contrast": null,
+  "is_prototype_wise_force_contrast": "1",
+  "is_prototype_wise_toallcelltype_contrast": 1,
+  "is_sample_wise": 1,
+  "layer_norm_eps": 1e-12,
+  "learnable_logits_scale": null,
+  "loss_weights": {
+    "cls_contrast_loss": 1.0,
+    "cls_prot_contrast_loss": 1.0,
+    "cls_self_contrast_loss": 0.0,
+    "plm_loss": 1.0,
+    "prot_prot_contrast_loss": 0.0,
+    "prot_self_contrast_loss": 0.0
+  },
+  "mask_expnum": true,
+  "max_length": null,
+  "max_position_embeddings": 2048,
+  "min_position": null,
+  "model_type": "bert",
+  "moving_avg_ratio": 0.9,
+  "no_l2norm": null,
+  "normalize_log1p": 1,
+  "num_attention_heads": 4,
+  "num_hidden_layers": 6,
+  "output_hidden_states": true,
+  "pad_token_id": 0,
+  "position_embedding_type": "absolute",
+  "project_linear_relu_linear_ln": null,
+  "prototype_scatter_type": "mixed_mean",
+  "prototype_type": "fully_learnable",
+  "randflip": null,
+  "randperm": null,
+  "rank_contrast_logits_clamp": null,
+  "rank_contrast_loss_type": "relative_softmax",
+  "relative_softmax_lossavg_bugfix": 1,
+  "relax_last_layernorm": 1,
+  "scale_grad_by_freq": true,
+  "separate_cls": 1,
+  "tau": 0.01,
+  "tf_class_vocab_size": null,
+  "tf_superclass_vocab_size": null,
+  "tie_word_embeddings": 0,
+  "toallcelltype_bugfix": 1,
+  "torch_dtype": "float32",
+  "total_logging_steps": 40,
+  "transform_linear": null,
+  "transform_no_layernorm": null,
+  "transform_swap_ln_relu": null,
+  "transformers_version": "4.32.1",
+  "truth_use_decoder": null,
+  "type_vocab_size": 1,
+  "use_abspos_embed": true,
+  "use_cache": true,
+  "use_tf_target_hardmask": false,
+  "vocab_size": 25427
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2bf9f8f95715b65f594c488f0501aa55b01df435e615c5bf0f795a31eda32418
+size 44049097