Upload config.json
Browse files- config.json +6 -2
config.json
CHANGED
@@ -12,7 +12,7 @@
|
|
12 |
"hidden_size": 2048,
|
13 |
"initializer_range": 0.02,
|
14 |
"intermediate_size": 6144,
|
15 |
-
"max_position_embeddings":
|
16 |
"max_window_layers": 48,
|
17 |
"mlp_only_layers": [],
|
18 |
"model_type": "qwen3_moe",
|
@@ -25,7 +25,11 @@
|
|
25 |
"num_key_value_heads": 4,
|
26 |
"output_router_logits": false,
|
27 |
"rms_norm_eps": 1e-06,
|
28 |
-
|
|
|
|
|
|
|
|
|
29 |
"rope_theta": 1000000.0,
|
30 |
"router_aux_loss_coef": 0.001,
|
31 |
"sliding_window": null,
|
|
|
12 |
"hidden_size": 2048,
|
13 |
"initializer_range": 0.02,
|
14 |
"intermediate_size": 6144,
|
15 |
+
"max_position_embeddings": 131072,
|
16 |
"max_window_layers": 48,
|
17 |
"mlp_only_layers": [],
|
18 |
"model_type": "qwen3_moe",
|
|
|
25 |
"num_key_value_heads": 4,
|
26 |
"output_router_logits": false,
|
27 |
"rms_norm_eps": 1e-06,
|
28 |
+
"rope_scaling": {
|
29 |
+
"type": "yarn",
|
30 |
+
"factor": 4.0,
|
31 |
+
"original_max_position_embeddings": 32768
|
32 |
+
},
|
33 |
"rope_theta": 1000000.0,
|
34 |
"router_aux_loss_coef": 0.001,
|
35 |
"sliding_window": null,
|