初始化项目,由ModelHub XC社区提供模型

Model: kshitijthakkar/loggenix-moe-0.3B-A0.1B-smolagents-v0-grpo
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-10-15 14:01:17 +08:00
commit 8b94363175
11 changed files with 478 additions and 0 deletions

40
config.json Normal file
View File

@@ -0,0 +1,40 @@
{
"architectures": [
"Qwen3MoeForCausalLM"
],
"attention_bias": false,
"attention_dropout": 0.1,
"decoder_sparse_step": 1,
"dtype": "bfloat16",
"eos_token_id": 151645,
"head_dim": 64,
"hidden_act": "silu",
"hidden_size": 512,
"initializer_range": 0.02,
"intermediate_size": 1536,
"max_position_embeddings": 8192,
"max_window_layers": 12,
"mlp_only_layers": [],
"model_type": "qwen3_moe",
"moe_intermediate_size": 512,
"norm_topk_prob": true,
"num_attention_heads": 8,
"num_experts": 16,
"num_experts_per_tok": 2,
"num_hidden_layers": 12,
"num_key_value_heads": 2,
"output_router_logits": true,
"pad_token_id": 151645,
"rms_norm_eps": 1e-06,
"rope_parameters": {
"rope_theta": 1000000,
"rope_type": "default"
},
"router_aux_loss_coef": 0.001,
"sliding_window": null,
"tie_word_embeddings": false,
"transformers_version": "5.0.0.dev0",
"use_cache": true,
"use_sliding_window": false,
"vocab_size": 151936
}