Update config.json
Browse files- config.json +1 -7
config.json
CHANGED
|
@@ -11,7 +11,7 @@
|
|
| 11 |
"hidden_size": 4096,
|
| 12 |
"initializer_range": 0.02,
|
| 13 |
"intermediate_size": 14336,
|
| 14 |
-
"max_position_embeddings":
|
| 15 |
"model_type": "mixtral",
|
| 16 |
"num_attention_heads": 32,
|
| 17 |
"num_experts_per_tok": 2,
|
|
@@ -21,12 +21,6 @@
|
|
| 21 |
"output_router_logits": false,
|
| 22 |
"pretraining_tp": 1,
|
| 23 |
"rms_norm_eps": 1e-05,
|
| 24 |
-
"rope_scaling": {
|
| 25 |
-
"factor": 5.0,
|
| 26 |
-
"original_max_position_embeddings": 8192,
|
| 27 |
-
"type": "yarn",
|
| 28 |
-
"finetuned": "true"
|
| 29 |
-
},
|
| 30 |
"rope_theta": 500000.0,
|
| 31 |
"router_aux_loss_coef": 0.001,
|
| 32 |
"router_jitter_noise": 0.0,
|
|
|
|
| 11 |
"hidden_size": 4096,
|
| 12 |
"initializer_range": 0.02,
|
| 13 |
"intermediate_size": 14336,
|
| 14 |
+
"max_position_embeddings": 8192,
|
| 15 |
"model_type": "mixtral",
|
| 16 |
"num_attention_heads": 32,
|
| 17 |
"num_experts_per_tok": 2,
|
|
|
|
| 21 |
"output_router_logits": false,
|
| 22 |
"pretraining_tp": 1,
|
| 23 |
"rms_norm_eps": 1e-05,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 24 |
"rope_theta": 500000.0,
|
| 25 |
"router_aux_loss_coef": 0.001,
|
| 26 |
"router_jitter_noise": 0.0,
|