Delete config.json
Browse files- config.json +0 -85
config.json
DELETED
|
@@ -1,85 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"producer": {
|
| 3 |
-
"name": "modelopt",
|
| 4 |
-
"version": "0.33.1"
|
| 5 |
-
},
|
| 6 |
-
"architecture": "Qwen3ForCausalLM",
|
| 7 |
-
"dtype": "bfloat16",
|
| 8 |
-
"logits_dtype": "float16",
|
| 9 |
-
"num_hidden_layers": 40,
|
| 10 |
-
"num_attention_heads": 40,
|
| 11 |
-
"num_key_value_heads": 8,
|
| 12 |
-
"hidden_size": 5120,
|
| 13 |
-
"norm_epsilon": 1e-06,
|
| 14 |
-
"vocab_size": 151936,
|
| 15 |
-
"max_position_embeddings": 40960,
|
| 16 |
-
"hidden_act": "silu",
|
| 17 |
-
"use_parallel_embedding": true,
|
| 18 |
-
"embedding_sharding_dim": 0,
|
| 19 |
-
"head_size": 128,
|
| 20 |
-
"intermediate_size": 17408,
|
| 21 |
-
"position_embedding_type": "rope_gpt_neox",
|
| 22 |
-
"share_embedding_table": false,
|
| 23 |
-
"residual_mlp": false,
|
| 24 |
-
"bias": false,
|
| 25 |
-
"rotary_pct": 1.0,
|
| 26 |
-
"rank": 0,
|
| 27 |
-
"decoder": "qwen",
|
| 28 |
-
"rmsnorm": true,
|
| 29 |
-
"lm_head_bias": false,
|
| 30 |
-
"mlp_bias": false,
|
| 31 |
-
"attn_bias": false,
|
| 32 |
-
"rotary_base": 1000000,
|
| 33 |
-
"rotary_scaling": null,
|
| 34 |
-
"disable_weight_only_quant_plugin": false,
|
| 35 |
-
"num_labels": 1,
|
| 36 |
-
"use_logn_attn": false,
|
| 37 |
-
"mlp_only_layers": [],
|
| 38 |
-
"decoder_sparse_step": 1,
|
| 39 |
-
"moe": {
|
| 40 |
-
"num_experts": 0,
|
| 41 |
-
"shared_expert_intermediate_size": 0,
|
| 42 |
-
"top_k": 0,
|
| 43 |
-
"normalization_mode": 0,
|
| 44 |
-
"sparse_mixer_epsilon": 0.01,
|
| 45 |
-
"tp_mode": 0,
|
| 46 |
-
"device_limited_n_group": 0,
|
| 47 |
-
"device_limited_topk_group": 0,
|
| 48 |
-
"device_limited_routed_scaling_factor": 1.0
|
| 49 |
-
},
|
| 50 |
-
"runtime_defaults": null,
|
| 51 |
-
"mapping": {
|
| 52 |
-
"world_size": 1,
|
| 53 |
-
"gpus_per_node": 8,
|
| 54 |
-
"cp_size": 1,
|
| 55 |
-
"tp_size": 1,
|
| 56 |
-
"pp_size": 1,
|
| 57 |
-
"moe_tp_size": 1,
|
| 58 |
-
"moe_cluster_size": 1,
|
| 59 |
-
"moe_ep_size": 1,
|
| 60 |
-
"attn_tp_size": 1,
|
| 61 |
-
"attn_cp_size": 1,
|
| 62 |
-
"auto_parallel": false
|
| 63 |
-
},
|
| 64 |
-
"quantization": {
|
| 65 |
-
"quant_algo": "FP8",
|
| 66 |
-
"kv_cache_quant_algo": "FP8",
|
| 67 |
-
"group_size": 128,
|
| 68 |
-
"smoothquant_val": 0.5,
|
| 69 |
-
"clamp_val": null,
|
| 70 |
-
"use_meta_recipe": false,
|
| 71 |
-
"has_zero_point": false,
|
| 72 |
-
"pre_quant_scale": false,
|
| 73 |
-
"exclude_modules": [
|
| 74 |
-
"lm_head"
|
| 75 |
-
]
|
| 76 |
-
},
|
| 77 |
-
"qk_layernorm": false,
|
| 78 |
-
"rotary_embedding_dim": 128,
|
| 79 |
-
"seq_length": 8192,
|
| 80 |
-
"qwen_type": "qwen3",
|
| 81 |
-
"moe_intermediate_size": 0,
|
| 82 |
-
"moe_shared_expert_intermediate_size": 0,
|
| 83 |
-
"tie_word_embeddings": false,
|
| 84 |
-
"model_type": "qwen"
|
| 85 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|