upload
Browse files- config.json +2 -11
config.json
CHANGED
|
@@ -14,7 +14,7 @@
|
|
| 14 |
"ep_size": 1,
|
| 15 |
"first_k_dense_replace": 3,
|
| 16 |
"hidden_act": "silu",
|
| 17 |
-
"hidden_size":
|
| 18 |
"initializer_range": 0.02,
|
| 19 |
"intermediate_size": 18432,
|
| 20 |
"kv_lora_rank": 512,
|
|
@@ -28,21 +28,12 @@
|
|
| 28 |
"norm_topk_prob": true,
|
| 29 |
"num_attention_heads": 128,
|
| 30 |
"num_experts_per_tok": 8,
|
| 31 |
-
"num_hidden_layers":
|
| 32 |
"num_key_value_heads": 128,
|
| 33 |
"num_nextn_predict_layers": 1,
|
| 34 |
"q_lora_rank": 1536,
|
| 35 |
"qk_nope_head_dim": 128,
|
| 36 |
"qk_rope_head_dim": 64,
|
| 37 |
-
"quantization_config": {
|
| 38 |
-
"activation_scheme": "dynamic",
|
| 39 |
-
"fmt": "e4m3",
|
| 40 |
-
"quant_method": "fp8",
|
| 41 |
-
"weight_block_size": [
|
| 42 |
-
128,
|
| 43 |
-
128
|
| 44 |
-
]
|
| 45 |
-
},
|
| 46 |
"rms_norm_eps": 1e-06,
|
| 47 |
"rope_scaling": {
|
| 48 |
"beta_fast": 32,
|
|
|
|
| 14 |
"ep_size": 1,
|
| 15 |
"first_k_dense_replace": 3,
|
| 16 |
"hidden_act": "silu",
|
| 17 |
+
"hidden_size": 128,
|
| 18 |
"initializer_range": 0.02,
|
| 19 |
"intermediate_size": 18432,
|
| 20 |
"kv_lora_rank": 512,
|
|
|
|
| 28 |
"norm_topk_prob": true,
|
| 29 |
"num_attention_heads": 128,
|
| 30 |
"num_experts_per_tok": 8,
|
| 31 |
+
"num_hidden_layers": 3,
|
| 32 |
"num_key_value_heads": 128,
|
| 33 |
"num_nextn_predict_layers": 1,
|
| 34 |
"q_lora_rank": 1536,
|
| 35 |
"qk_nope_head_dim": 128,
|
| 36 |
"qk_rope_head_dim": 64,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 37 |
"rms_norm_eps": 1e-06,
|
| 38 |
"rope_scaling": {
|
| 39 |
"beta_fast": 32,
|