Upload config.json with huggingface_hub
Browse files- config.json +84 -0
config.json
ADDED
|
@@ -0,0 +1,84 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Qwen3MoeForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_bias": false,
|
| 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",
|
| 13 |
+
"hidden_size": 2048,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 6144,
|
| 16 |
+
"max_position_embeddings": 262144,
|
| 17 |
+
"max_window_layers": 48,
|
| 18 |
+
"mlp_only_layers": [],
|
| 19 |
+
"model_type": "qwen3_moe",
|
| 20 |
+
"moe_intermediate_size": 768,
|
| 21 |
+
"norm_topk_prob": true,
|
| 22 |
+
"num_attention_heads": 32,
|
| 23 |
+
"num_experts": 128,
|
| 24 |
+
"num_experts_per_tok": 8,
|
| 25 |
+
"num_hidden_layers": 48,
|
| 26 |
+
"num_key_value_heads": 4,
|
| 27 |
+
"output_router_logits": false,
|
| 28 |
+
"quantization_config": {
|
| 29 |
+
"config_groups": {
|
| 30 |
+
"group_0": {
|
| 31 |
+
"format": "float-quantized",
|
| 32 |
+
"input_activations": {
|
| 33 |
+
"actorder": null,
|
| 34 |
+
"block_structure": null,
|
| 35 |
+
"dynamic": true,
|
| 36 |
+
"group_size": null,
|
| 37 |
+
"num_bits": 8,
|
| 38 |
+
"observer": null,
|
| 39 |
+
"observer_kwargs": {},
|
| 40 |
+
"strategy": "token",
|
| 41 |
+
"symmetric": true,
|
| 42 |
+
"type": "float"
|
| 43 |
+
},
|
| 44 |
+
"output_activations": null,
|
| 45 |
+
"targets": [
|
| 46 |
+
"Linear"
|
| 47 |
+
],
|
| 48 |
+
"weights": {
|
| 49 |
+
"actorder": null,
|
| 50 |
+
"block_structure": null,
|
| 51 |
+
"dynamic": false,
|
| 52 |
+
"group_size": null,
|
| 53 |
+
"num_bits": 8,
|
| 54 |
+
"observer": "minmax",
|
| 55 |
+
"observer_kwargs": {},
|
| 56 |
+
"strategy": "channel",
|
| 57 |
+
"symmetric": true,
|
| 58 |
+
"type": "float"
|
| 59 |
+
}
|
| 60 |
+
}
|
| 61 |
+
},
|
| 62 |
+
"format": "float-quantized",
|
| 63 |
+
"global_compression_ratio": null,
|
| 64 |
+
"ignore": [
|
| 65 |
+
"lm_head"
|
| 66 |
+
],
|
| 67 |
+
"kv_cache_scheme": null,
|
| 68 |
+
"quant_method": "compressed-tensors",
|
| 69 |
+
"quantization_status": "compressed",
|
| 70 |
+
"sparsity_config": {},
|
| 71 |
+
"transform_config": {},
|
| 72 |
+
"version": "0.12.2"
|
| 73 |
+
},
|
| 74 |
+
"rms_norm_eps": 1e-06,
|
| 75 |
+
"rope_scaling": null,
|
| 76 |
+
"rope_theta": 10000000,
|
| 77 |
+
"router_aux_loss_coef": 0.001,
|
| 78 |
+
"sliding_window": null,
|
| 79 |
+
"tie_word_embeddings": false,
|
| 80 |
+
"transformers_version": "4.56.2",
|
| 81 |
+
"use_cache": true,
|
| 82 |
+
"use_sliding_window": false,
|
| 83 |
+
"vocab_size": 151936
|
| 84 |
+
}
|