jsunn-y
commited on
Commit
•
e9e73af
1
Parent(s):
5e64795
adding new text-guided model
Browse files
text-proteinDT-swissprot/1.5B/config.json
ADDED
|
@@ -0,0 +1,67 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_function": "gelu_new",
|
| 3 |
+
"adapter_c_hidden": 16,
|
| 4 |
+
"adapter_c_s": 256,
|
| 5 |
+
"adapter_dropout": 0.1,
|
| 6 |
+
"adapter_nlayers": 4,
|
| 7 |
+
"adapter_projection_nlayers": 2,
|
| 8 |
+
"adapter_shared_projection": true,
|
| 9 |
+
"adapter_summation": false,
|
| 10 |
+
"adapter_weight_init": 1e-05,
|
| 11 |
+
"architectures": [
|
| 12 |
+
"ProgenConditional"
|
| 13 |
+
],
|
| 14 |
+
"attn_pdrop": 0.0,
|
| 15 |
+
"bos_token_id": 1,
|
| 16 |
+
"conditions_shared_adapter": true,
|
| 17 |
+
"embd_pdrop": 0.0,
|
| 18 |
+
"encoding_dimensions": {
|
| 19 |
+
"text": 256
|
| 20 |
+
},
|
| 21 |
+
"encoding_files": {
|
| 22 |
+
"text": "data/text2encoding.pt"
|
| 23 |
+
},
|
| 24 |
+
"eos_token_id": 2,
|
| 25 |
+
"full_finetuning": false,
|
| 26 |
+
"full_fineturning": false,
|
| 27 |
+
"gradient_checkpointing": true,
|
| 28 |
+
"initializer_range": 0.02,
|
| 29 |
+
"inter_context_sequence_attention": true,
|
| 30 |
+
"inter_sequence_attention": true,
|
| 31 |
+
"layer_norm_epsilon": 1e-05,
|
| 32 |
+
"max_num_sequences": 512,
|
| 33 |
+
"model_type": "progen-conditional",
|
| 34 |
+
"msa_style_attention": true,
|
| 35 |
+
"n_ctx": 2048,
|
| 36 |
+
"n_embd": 1536,
|
| 37 |
+
"n_head": 16,
|
| 38 |
+
"n_inner": null,
|
| 39 |
+
"n_layer": 27,
|
| 40 |
+
"n_positions": 2048,
|
| 41 |
+
"output_router_logits": true,
|
| 42 |
+
"pad_token_id": 0,
|
| 43 |
+
"pretrained_model_dir": "data/pretrained_models/progen2-base",
|
| 44 |
+
"pretrained_model_name": "progen2-base",
|
| 45 |
+
"resid_pdrop": 0.0,
|
| 46 |
+
"rotary_dim": 48,
|
| 47 |
+
"router_aux_loss_coef": 0.05,
|
| 48 |
+
"scale_attn_weights": true,
|
| 49 |
+
"summary_activation": null,
|
| 50 |
+
"summary_first_dropout": 0.1,
|
| 51 |
+
"summary_proj_to_labels": true,
|
| 52 |
+
"summary_type": "cls_index",
|
| 53 |
+
"summary_use_proj": true,
|
| 54 |
+
"task_specific_params": {
|
| 55 |
+
"text-generation": {
|
| 56 |
+
"do_sample": true,
|
| 57 |
+
"max_length": 50,
|
| 58 |
+
"temperature": 1.0
|
| 59 |
+
}
|
| 60 |
+
},
|
| 61 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 62 |
+
"torch_dtype": "bfloat16",
|
| 63 |
+
"transformers_version": "4.46.2",
|
| 64 |
+
"unified_active_context_pos_ids": true,
|
| 65 |
+
"use_cache": true,
|
| 66 |
+
"vocab_size": 32
|
| 67 |
+
}
|
text-proteinDT-swissprot/1.5B/generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"pad_token_id": 0,
|
| 6 |
+
"transformers_version": "4.46.2"
|
| 7 |
+
}
|
text-proteinDT-swissprot/1.5B/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bf780bdce8e50784eea8fd3962e34fbd01143008fc04a04b915d59eb26aac51a
|
| 3 |
+
size 1828528694
|