Upload folder using huggingface_hub
Browse files- config.json +6 -9
- generation_config.json +1 -1
config.json
CHANGED
|
@@ -6,7 +6,6 @@
|
|
| 6 |
"attention_dropout": 0.0,
|
| 7 |
"bos_token_id": 151643,
|
| 8 |
"decoder_sparse_step": 1,
|
| 9 |
-
"dtype": "bfloat16",
|
| 10 |
"eos_token_id": 151645,
|
| 11 |
"head_dim": 128,
|
| 12 |
"hidden_act": "silu",
|
|
@@ -20,12 +19,11 @@
|
|
| 20 |
"moe_intermediate_size": 768,
|
| 21 |
"norm_topk_prob": true,
|
| 22 |
"num_attention_heads": 32,
|
|
|
|
| 23 |
"num_experts_per_tok": 8,
|
| 24 |
"num_hidden_layers": 48,
|
| 25 |
"num_key_value_heads": 4,
|
| 26 |
-
"num_local_experts": 128,
|
| 27 |
"output_router_logits": false,
|
| 28 |
-
"pad_token_id": null,
|
| 29 |
"quantization_config": {
|
| 30 |
"config_groups": {
|
| 31 |
"group_0": {
|
|
@@ -122,17 +120,16 @@
|
|
| 122 |
"quantization_status": "compressed",
|
| 123 |
"sparsity_config": {},
|
| 124 |
"transform_config": {},
|
| 125 |
-
"version": "0.18.1.
|
| 126 |
},
|
| 127 |
"rms_norm_eps": 1e-06,
|
| 128 |
-
"
|
| 129 |
-
|
| 130 |
-
"rope_type": "default"
|
| 131 |
-
},
|
| 132 |
"router_aux_loss_coef": 0.001,
|
| 133 |
"sliding_window": null,
|
| 134 |
"tie_word_embeddings": false,
|
| 135 |
-
"
|
|
|
|
| 136 |
"use_cache": true,
|
| 137 |
"use_sliding_window": false,
|
| 138 |
"vocab_size": 151936
|
|
|
|
| 6 |
"attention_dropout": 0.0,
|
| 7 |
"bos_token_id": 151643,
|
| 8 |
"decoder_sparse_step": 1,
|
|
|
|
| 9 |
"eos_token_id": 151645,
|
| 10 |
"head_dim": 128,
|
| 11 |
"hidden_act": "silu",
|
|
|
|
| 19 |
"moe_intermediate_size": 768,
|
| 20 |
"norm_topk_prob": true,
|
| 21 |
"num_attention_heads": 32,
|
| 22 |
+
"num_experts": 128,
|
| 23 |
"num_experts_per_tok": 8,
|
| 24 |
"num_hidden_layers": 48,
|
| 25 |
"num_key_value_heads": 4,
|
|
|
|
| 26 |
"output_router_logits": false,
|
|
|
|
| 27 |
"quantization_config": {
|
| 28 |
"config_groups": {
|
| 29 |
"group_0": {
|
|
|
|
| 120 |
"quantization_status": "compressed",
|
| 121 |
"sparsity_config": {},
|
| 122 |
"transform_config": {},
|
| 123 |
+
"version": "0.18.1.a20260910"
|
| 124 |
},
|
| 125 |
"rms_norm_eps": 1e-06,
|
| 126 |
+
"rope_scaling": null,
|
| 127 |
+
"rope_theta": 1000000.0,
|
|
|
|
|
|
|
| 128 |
"router_aux_loss_coef": 0.001,
|
| 129 |
"sliding_window": null,
|
| 130 |
"tie_word_embeddings": false,
|
| 131 |
+
"torch_dtype": "bfloat16",
|
| 132 |
+
"transformers_version": "4.51.0",
|
| 133 |
"use_cache": true,
|
| 134 |
"use_sliding_window": false,
|
| 135 |
"vocab_size": 151936
|
generation_config.json
CHANGED
|
@@ -9,5 +9,5 @@
|
|
| 9 |
"temperature": 0.6,
|
| 10 |
"top_k": 20,
|
| 11 |
"top_p": 0.95,
|
| 12 |
-
"transformers_version": "5.
|
| 13 |
}
|
|
|
|
| 9 |
"temperature": 0.6,
|
| 10 |
"top_k": 20,
|
| 11 |
"top_p": 0.95,
|
| 12 |
+
"transformers_version": "5.17.0"
|
| 13 |
}
|