2025-02-14_specter2_contrastive_fold_0
Browse files- config.json +57 -15
- model.safetensors +2 -2
config.json
CHANGED
|
@@ -1,28 +1,70 @@
|
|
| 1 |
{
|
| 2 |
-
"_name_or_path": "
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
"architectures": [
|
| 4 |
"CustomModel"
|
| 5 |
],
|
| 6 |
"attention_probs_dropout_prob": 0.1,
|
| 7 |
-
"bos_token_id": 0,
|
| 8 |
"classifier_dropout": null,
|
| 9 |
-
"eos_token_id": 2,
|
| 10 |
"hidden_act": "gelu",
|
| 11 |
"hidden_dropout_prob": 0.1,
|
| 12 |
-
"hidden_size":
|
| 13 |
"initializer_range": 0.02,
|
| 14 |
-
"intermediate_size":
|
| 15 |
-
"layer_norm_eps": 1e-
|
| 16 |
-
"max_position_embeddings":
|
| 17 |
-
"model_type": "
|
| 18 |
-
"num_attention_heads":
|
| 19 |
-
"num_hidden_layers":
|
| 20 |
-
"
|
| 21 |
-
"pad_token_id": 1,
|
| 22 |
"position_embedding_type": "absolute",
|
| 23 |
-
"
|
|
|
|
| 24 |
"transformers_version": "4.47.1",
|
| 25 |
-
"type_vocab_size":
|
| 26 |
"use_cache": true,
|
| 27 |
-
"vocab_size":
|
| 28 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "specter2",
|
| 3 |
+
"adapters": {
|
| 4 |
+
"adapters": {
|
| 5 |
+
"specter2": "9076f36a74755ac4"
|
| 6 |
+
},
|
| 7 |
+
"config_map": {
|
| 8 |
+
"9076f36a74755ac4": {
|
| 9 |
+
"adapter_residual_before_ln": false,
|
| 10 |
+
"cross_adapter": false,
|
| 11 |
+
"dropout": 0.0,
|
| 12 |
+
"factorized_phm_W": true,
|
| 13 |
+
"factorized_phm_rule": false,
|
| 14 |
+
"hypercomplex_nonlinearity": "glorot-uniform",
|
| 15 |
+
"init_weights": "bert",
|
| 16 |
+
"inv_adapter": null,
|
| 17 |
+
"inv_adapter_reduction_factor": null,
|
| 18 |
+
"is_parallel": false,
|
| 19 |
+
"learn_phm": true,
|
| 20 |
+
"leave_out": [],
|
| 21 |
+
"ln_after": false,
|
| 22 |
+
"ln_before": false,
|
| 23 |
+
"mh_adapter": false,
|
| 24 |
+
"non_linearity": "relu",
|
| 25 |
+
"original_ln_after": true,
|
| 26 |
+
"original_ln_before": true,
|
| 27 |
+
"output_adapter": true,
|
| 28 |
+
"phm_bias": true,
|
| 29 |
+
"phm_c_init": "normal",
|
| 30 |
+
"phm_dim": 4,
|
| 31 |
+
"phm_init_range": 0.0001,
|
| 32 |
+
"phm_layer": false,
|
| 33 |
+
"phm_rank": 1,
|
| 34 |
+
"reduction_factor": 16,
|
| 35 |
+
"residual_before_ln": true,
|
| 36 |
+
"scaling": 1.0,
|
| 37 |
+
"shared_W_phm": false,
|
| 38 |
+
"shared_phm_rule": true,
|
| 39 |
+
"stochastic_depth": 0.0,
|
| 40 |
+
"use_gating": false
|
| 41 |
+
}
|
| 42 |
+
},
|
| 43 |
+
"fusion_config_map": {},
|
| 44 |
+
"fusion_name_map": {},
|
| 45 |
+
"fusions": {}
|
| 46 |
+
},
|
| 47 |
"architectures": [
|
| 48 |
"CustomModel"
|
| 49 |
],
|
| 50 |
"attention_probs_dropout_prob": 0.1,
|
|
|
|
| 51 |
"classifier_dropout": null,
|
|
|
|
| 52 |
"hidden_act": "gelu",
|
| 53 |
"hidden_dropout_prob": 0.1,
|
| 54 |
+
"hidden_size": 768,
|
| 55 |
"initializer_range": 0.02,
|
| 56 |
+
"intermediate_size": 3072,
|
| 57 |
+
"layer_norm_eps": 1e-12,
|
| 58 |
+
"max_position_embeddings": 512,
|
| 59 |
+
"model_type": "bert",
|
| 60 |
+
"num_attention_heads": 12,
|
| 61 |
+
"num_hidden_layers": 12,
|
| 62 |
+
"pad_token_id": 0,
|
|
|
|
| 63 |
"position_embedding_type": "absolute",
|
| 64 |
+
"prediction_heads": {},
|
| 65 |
+
"torch_dtype": "float32",
|
| 66 |
"transformers_version": "4.47.1",
|
| 67 |
+
"type_vocab_size": 2,
|
| 68 |
"use_cache": true,
|
| 69 |
+
"vocab_size": 31096
|
| 70 |
}
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:81651fc08ba418def3d6134995e255709b2239e8c5ce2a52fbceb19196637e5f
|
| 3 |
+
size 571522000
|