Delete config.json
Browse files- config.json +0 -95
config.json
DELETED
|
@@ -1,95 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"model_type": "blt",
|
| 3 |
-
"vocab_size": 260,
|
| 4 |
-
"max_position_embeddings": 4096,
|
| 5 |
-
"patch_in_forward": true,
|
| 6 |
-
"realtime_patching": true,
|
| 7 |
-
"patching_mode": "entropy",
|
| 8 |
-
"patch_size": 4,
|
| 9 |
-
"patching_threshold": 1.335442066192627,
|
| 10 |
-
"patching_threshold_add": null,
|
| 11 |
-
"max_patch_length": null,
|
| 12 |
-
"patching_batch_size": 1,
|
| 13 |
-
"patching_device": "cuda",
|
| 14 |
-
"monotonicity": false,
|
| 15 |
-
"cross_attn_k": 2,
|
| 16 |
-
"encoder_hash_byte_group_size": [
|
| 17 |
-
3,
|
| 18 |
-
4,
|
| 19 |
-
5,
|
| 20 |
-
6,
|
| 21 |
-
7,
|
| 22 |
-
8
|
| 23 |
-
],
|
| 24 |
-
"encoder_hash_byte_group_vocab": 500002,
|
| 25 |
-
"encoder_hash_byte_group_nb_functions": 1,
|
| 26 |
-
"pm_size": 0,
|
| 27 |
-
"patcher_config": {
|
| 28 |
-
"vocab_size": 260,
|
| 29 |
-
"hidden_size": 768,
|
| 30 |
-
"num_hidden_layers": 14,
|
| 31 |
-
"num_attention_heads": 12,
|
| 32 |
-
"num_key_value_heads": null,
|
| 33 |
-
"max_position_embeddings": 8192,
|
| 34 |
-
"norm_eps": 1e-05,
|
| 35 |
-
"dropout": 0.0,
|
| 36 |
-
"rope_theta": 10000.0,
|
| 37 |
-
"attn_bias_type": "local_block_causal",
|
| 38 |
-
"intermediate_size": 2048
|
| 39 |
-
},
|
| 40 |
-
"encoder_config": {
|
| 41 |
-
"vocab_size": 260,
|
| 42 |
-
"cross_attn_all_layers": false,
|
| 43 |
-
"cross_attn_k": 2,
|
| 44 |
-
"hidden_size_global": 2048,
|
| 45 |
-
"pm_size": 0,
|
| 46 |
-
"hidden_size": 1024,
|
| 47 |
-
"num_attention_heads": 16,
|
| 48 |
-
"num_key_value_heads": null,
|
| 49 |
-
"num_hidden_layers": 1,
|
| 50 |
-
"norm_eps": 1e-05,
|
| 51 |
-
"dropout": 0.0,
|
| 52 |
-
"max_position_embeddings": 24576,
|
| 53 |
-
"rope_theta": 500000.0,
|
| 54 |
-
"rope_scaling": {
|
| 55 |
-
"rope_type": "default"
|
| 56 |
-
},
|
| 57 |
-
"hidden_act": "silu",
|
| 58 |
-
"intermediate_size": 2816
|
| 59 |
-
},
|
| 60 |
-
"decoder_config": {
|
| 61 |
-
"vocab_size": 260,
|
| 62 |
-
"cross_attn_all_layers": true,
|
| 63 |
-
"cross_attn_k": 2,
|
| 64 |
-
"hidden_size_global": 2048,
|
| 65 |
-
"hidden_size": 1024,
|
| 66 |
-
"num_attention_heads": 16,
|
| 67 |
-
"num_key_value_heads": null,
|
| 68 |
-
"num_hidden_layers": 9,
|
| 69 |
-
"norm_eps": 1e-05,
|
| 70 |
-
"dropout": 0.0,
|
| 71 |
-
"max_position_embeddings": 24576,
|
| 72 |
-
"rope_theta": 500000.0,
|
| 73 |
-
"rope_scaling": {
|
| 74 |
-
"rope_type": "default"
|
| 75 |
-
},
|
| 76 |
-
"hidden_act": "silu",
|
| 77 |
-
"intermediate_size": 2816
|
| 78 |
-
},
|
| 79 |
-
"global_config": {
|
| 80 |
-
"hidden_size": 2048,
|
| 81 |
-
"num_attention_heads": 16,
|
| 82 |
-
"num_key_value_heads": null,
|
| 83 |
-
"num_hidden_layers": 25,
|
| 84 |
-
"norm_eps": 1e-05,
|
| 85 |
-
"dropout": 0.0,
|
| 86 |
-
"max_position_embeddings": 4096,
|
| 87 |
-
"rope_theta": 500000.0,
|
| 88 |
-
"rope_scaling": {
|
| 89 |
-
"rope_type": "default"
|
| 90 |
-
},
|
| 91 |
-
"hidden_act": "silu",
|
| 92 |
-
"intermediate_size": 5632
|
| 93 |
-
},
|
| 94 |
-
"tie_word_embeddings": false
|
| 95 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|