RthItalia/NanoLLM-Qwen2.5-14B-v3.1
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

初始化项目，由ModelHub XC社区提供模型

Browse Source 
Model: RthItalia/NanoLLM-Qwen2.5-14B-v3.1
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-05-09 10:59:03 +08:00
commit d652fe3c32
20 changed files with 457898 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

27
spec.json Normal file
View File

@@ -0,0 +1,27 @@
{
"format": "nano-v3.1-multi",
"base_model_id": "Qwen/Qwen2.5-14B-Instruct",
"hidden_layers": 48,
"locked_count": 76,
"pending_8bit": 260,
"elapsed_seconds": 5959,
"build_reference_mode": "8bit",
"reference_scope": "original_baseline",
"pending_policy": "leave_in_base_8bit",
"pruned_from": "/kaggle/working/final_artifact_Qwen2.5-14B-Instruct",
"pruned_strategy": "remove_last_locked",
"pruned_removed_count": 9,
"pruned_removed_modules": [
"model.layers.42.mlp.down_proj",
"model.layers.45.mlp.gate_proj",
"model.layers.46.mlp.down_proj",
"model.layers.28.self_attn.v_proj",
"model.layers.29.mlp.down_proj",
"model.layers.34.mlp.gate_proj",
"model.layers.36.mlp.up_proj",
"model.layers.42.self_attn.k_proj",
"model.layers.46.self_attn.o_proj"
],
"self_contained": true,
"base_model_local_subdir": "."
}