add rope_scaling to model config
Browse files- config.json +6 -2
config.json
CHANGED
|
@@ -13,7 +13,12 @@
|
|
| 13 |
"hidden_size": 8192,
|
| 14 |
"initializer_range": 0.02,
|
| 15 |
"intermediate_size": 28672,
|
| 16 |
-
"max_position_embeddings":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 17 |
"mlp_bias": false,
|
| 18 |
"model_type": "llama",
|
| 19 |
"num_attention_heads": 64,
|
|
@@ -24,7 +29,6 @@
|
|
| 24 |
"pretraining_tp": 1,
|
| 25 |
"resid_pdrop": 0.0,
|
| 26 |
"rms_norm_eps": 1e-05,
|
| 27 |
-
"rope_scaling": null,
|
| 28 |
"rope_theta": 10000000,
|
| 29 |
"tie_word_embeddings": false,
|
| 30 |
"transformers_version": "4.57.1",
|
|
|
|
| 13 |
"hidden_size": 8192,
|
| 14 |
"initializer_range": 0.02,
|
| 15 |
"intermediate_size": 28672,
|
| 16 |
+
"max_position_embeddings": 262144,
|
| 17 |
+
"rope_scaling": {
|
| 18 |
+
"type": "yarn",
|
| 19 |
+
"factor": 2.0,
|
| 20 |
+
"original_max_position_embeddings": 131072
|
| 21 |
+
},
|
| 22 |
"mlp_bias": false,
|
| 23 |
"model_type": "llama",
|
| 24 |
"num_attention_heads": 64,
|
|
|
|
| 29 |
"pretraining_tp": 1,
|
| 30 |
"resid_pdrop": 0.0,
|
| 31 |
"rms_norm_eps": 1e-05,
|
|
|
|
| 32 |
"rope_theta": 10000000,
|
| 33 |
"tie_word_embeddings": false,
|
| 34 |
"transformers_version": "4.57.1",
|