Update model/config.json
Browse files- model/config.json +16 -5
model/config.json
CHANGED
|
@@ -8,7 +8,10 @@
|
|
| 8 |
"attn_logit_softcapping": null,
|
| 9 |
"bos_token_id": 2,
|
| 10 |
"dtype": "bfloat16",
|
| 11 |
-
"eos_token_id":
|
|
|
|
|
|
|
|
|
|
| 12 |
"final_logit_softcapping": null,
|
| 13 |
"head_dim": 256,
|
| 14 |
"hidden_activation": "gelu_pytorch_tanh",
|
|
@@ -43,11 +46,19 @@
|
|
| 43 |
"pad_token_id": 0,
|
| 44 |
"query_pre_attn_scalar": 256,
|
| 45 |
"rms_norm_eps": 1e-06,
|
| 46 |
-
"
|
| 47 |
-
|
| 48 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 49 |
"sliding_window": 512,
|
| 50 |
-
"
|
|
|
|
| 51 |
"use_bidirectional_attention": false,
|
| 52 |
"use_cache": true,
|
| 53 |
"vocab_size": 262144
|
|
|
|
| 8 |
"attn_logit_softcapping": null,
|
| 9 |
"bos_token_id": 2,
|
| 10 |
"dtype": "bfloat16",
|
| 11 |
+
"eos_token_id": [
|
| 12 |
+
1,
|
| 13 |
+
50
|
| 14 |
+
],
|
| 15 |
"final_logit_softcapping": null,
|
| 16 |
"head_dim": 256,
|
| 17 |
"hidden_activation": "gelu_pytorch_tanh",
|
|
|
|
| 46 |
"pad_token_id": 0,
|
| 47 |
"query_pre_attn_scalar": 256,
|
| 48 |
"rms_norm_eps": 1e-06,
|
| 49 |
+
"rope_parameters": {
|
| 50 |
+
"full_attention": {
|
| 51 |
+
"rope_theta": 1000000.0,
|
| 52 |
+
"rope_type": "default"
|
| 53 |
+
},
|
| 54 |
+
"sliding_attention": {
|
| 55 |
+
"rope_theta": 10000.0,
|
| 56 |
+
"rope_type": "default"
|
| 57 |
+
}
|
| 58 |
+
},
|
| 59 |
"sliding_window": 512,
|
| 60 |
+
"tie_word_embeddings": true,
|
| 61 |
+
"transformers_version": "5.2.0",
|
| 62 |
"use_bidirectional_attention": false,
|
| 63 |
"use_cache": true,
|
| 64 |
"vocab_size": 262144
|