Upload 4 files
Browse files- README.md +7 -0
- adapter_config.json +33 -0
- adapter_model.safetensors +3 -0
- nas_lora_config.json +1 -0
README.md
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: meta-llama/Llama-3.2-11B-Vision-Instruct
|
| 3 |
+
library_name: peft
|
| 4 |
+
---
|
| 5 |
+
### Framework versions
|
| 6 |
+
|
| 7 |
+
- PEFT 0.13.2
|
adapter_config.json
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alpha_pattern": {},
|
| 3 |
+
"auto_mapping": null,
|
| 4 |
+
"base_model_name_or_path": "meta-llama/Llama-3.2-11B-Vision-Instruct",
|
| 5 |
+
"bias": "none",
|
| 6 |
+
"eva_config": null,
|
| 7 |
+
"exclude_modules": null,
|
| 8 |
+
"fan_in_fan_out": false,
|
| 9 |
+
"inference_mode": true,
|
| 10 |
+
"init_lora_weights": true,
|
| 11 |
+
"layer_replication": null,
|
| 12 |
+
"layers_pattern": null,
|
| 13 |
+
"layers_to_transform": null,
|
| 14 |
+
"loftq_config": {},
|
| 15 |
+
"lora_alpha": 32,
|
| 16 |
+
"lora_bias": false,
|
| 17 |
+
"lora_dropout": 0.05,
|
| 18 |
+
"megatron_config": null,
|
| 19 |
+
"megatron_core": "megatron.core",
|
| 20 |
+
"modules_to_save": null,
|
| 21 |
+
"peft_type": "LORA",
|
| 22 |
+
"r": 64,
|
| 23 |
+
"rank_pattern": {},
|
| 24 |
+
"revision": null,
|
| 25 |
+
"target_modules": [
|
| 26 |
+
"gate_proj",
|
| 27 |
+
"up_proj",
|
| 28 |
+
"down_proj"
|
| 29 |
+
],
|
| 30 |
+
"task_type": "CAUSAL_LM",
|
| 31 |
+
"use_dora": false,
|
| 32 |
+
"use_rslora": false
|
| 33 |
+
}
|
adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:35da050eac6e22bffde4abf8baa71ffa821d2a108e4410d488ed314be6d5f7c7
|
| 3 |
+
size 246581800
|
nas_lora_config.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"base_model.model.language_model.model.layers.0.mlp.gate_proj": 8, "base_model.model.language_model.model.layers.0.mlp.up_proj": 32, "base_model.model.language_model.model.layers.0.mlp.down_proj": 8, "base_model.model.language_model.model.layers.1.mlp.gate_proj": 64, "base_model.model.language_model.model.layers.1.mlp.up_proj": 16, "base_model.model.language_model.model.layers.1.mlp.down_proj": 16, "base_model.model.language_model.model.layers.2.mlp.gate_proj": 64, "base_model.model.language_model.model.layers.2.mlp.up_proj": 32, "base_model.model.language_model.model.layers.2.mlp.down_proj": 64, "base_model.model.language_model.model.layers.3.mlp.gate_proj": 16, "base_model.model.language_model.model.layers.3.mlp.up_proj": 16, "base_model.model.language_model.model.layers.3.mlp.down_proj": 16, "base_model.model.language_model.model.layers.4.mlp.gate_proj": 64, "base_model.model.language_model.model.layers.4.mlp.up_proj": 64, "base_model.model.language_model.model.layers.4.mlp.down_proj": 32, "base_model.model.language_model.model.layers.5.mlp.gate_proj": 8, "base_model.model.language_model.model.layers.5.mlp.up_proj": 64, "base_model.model.language_model.model.layers.5.mlp.down_proj": 32, "base_model.model.language_model.model.layers.6.mlp.gate_proj": 64, "base_model.model.language_model.model.layers.6.mlp.up_proj": 8, "base_model.model.language_model.model.layers.6.mlp.down_proj": 4, "base_model.model.language_model.model.layers.7.mlp.gate_proj": 32, "base_model.model.language_model.model.layers.7.mlp.up_proj": 64, "base_model.model.language_model.model.layers.7.mlp.down_proj": 16, "base_model.model.language_model.model.layers.8.mlp.gate_proj": 8, "base_model.model.language_model.model.layers.8.mlp.up_proj": 4, "base_model.model.language_model.model.layers.8.mlp.down_proj": 64, "base_model.model.language_model.model.layers.9.mlp.gate_proj": 4, "base_model.model.language_model.model.layers.9.mlp.up_proj": 64, "base_model.model.language_model.model.layers.9.mlp.down_proj": 8, "base_model.model.language_model.model.layers.10.mlp.gate_proj": 32, "base_model.model.language_model.model.layers.10.mlp.up_proj": 16, "base_model.model.language_model.model.layers.10.mlp.down_proj": 32, "base_model.model.language_model.model.layers.11.mlp.gate_proj": 16, "base_model.model.language_model.model.layers.11.mlp.up_proj": 32, "base_model.model.language_model.model.layers.11.mlp.down_proj": 4, "base_model.model.language_model.model.layers.12.mlp.gate_proj": 32, "base_model.model.language_model.model.layers.12.mlp.up_proj": 16, "base_model.model.language_model.model.layers.12.mlp.down_proj": 4, "base_model.model.language_model.model.layers.13.mlp.gate_proj": 32, "base_model.model.language_model.model.layers.13.mlp.up_proj": 32, "base_model.model.language_model.model.layers.13.mlp.down_proj": 64, "base_model.model.language_model.model.layers.14.mlp.gate_proj": 8, "base_model.model.language_model.model.layers.14.mlp.up_proj": 8, "base_model.model.language_model.model.layers.14.mlp.down_proj": 32, "base_model.model.language_model.model.layers.15.mlp.gate_proj": 16, "base_model.model.language_model.model.layers.15.mlp.up_proj": 16, "base_model.model.language_model.model.layers.15.mlp.down_proj": 64, "base_model.model.language_model.model.layers.16.mlp.gate_proj": 64, "base_model.model.language_model.model.layers.16.mlp.up_proj": 16, "base_model.model.language_model.model.layers.16.mlp.down_proj": 32, "base_model.model.language_model.model.layers.17.mlp.gate_proj": 8, "base_model.model.language_model.model.layers.17.mlp.up_proj": 64, "base_model.model.language_model.model.layers.17.mlp.down_proj": 4, "base_model.model.language_model.model.layers.18.mlp.gate_proj": 64, "base_model.model.language_model.model.layers.18.mlp.up_proj": 8, "base_model.model.language_model.model.layers.18.mlp.down_proj": 4, "base_model.model.language_model.model.layers.19.mlp.gate_proj": 64, "base_model.model.language_model.model.layers.19.mlp.up_proj": 16, "base_model.model.language_model.model.layers.19.mlp.down_proj": 8, "base_model.model.language_model.model.layers.20.mlp.gate_proj": 8, "base_model.model.language_model.model.layers.20.mlp.up_proj": 16, "base_model.model.language_model.model.layers.20.mlp.down_proj": 32, "base_model.model.language_model.model.layers.21.mlp.gate_proj": 4, "base_model.model.language_model.model.layers.21.mlp.up_proj": 8, "base_model.model.language_model.model.layers.21.mlp.down_proj": 4, "base_model.model.language_model.model.layers.22.mlp.gate_proj": 4, "base_model.model.language_model.model.layers.22.mlp.up_proj": 4, "base_model.model.language_model.model.layers.22.mlp.down_proj": 8, "base_model.model.language_model.model.layers.23.mlp.gate_proj": 32, "base_model.model.language_model.model.layers.23.mlp.up_proj": 32, "base_model.model.language_model.model.layers.23.mlp.down_proj": 8, "base_model.model.language_model.model.layers.24.mlp.gate_proj": 64, "base_model.model.language_model.model.layers.24.mlp.up_proj": 16, "base_model.model.language_model.model.layers.24.mlp.down_proj": 64, "base_model.model.language_model.model.layers.25.mlp.gate_proj": 8, "base_model.model.language_model.model.layers.25.mlp.up_proj": 32, "base_model.model.language_model.model.layers.25.mlp.down_proj": 32, "base_model.model.language_model.model.layers.26.mlp.gate_proj": 8, "base_model.model.language_model.model.layers.26.mlp.up_proj": 32, "base_model.model.language_model.model.layers.26.mlp.down_proj": 32, "base_model.model.language_model.model.layers.27.mlp.gate_proj": 8, "base_model.model.language_model.model.layers.27.mlp.up_proj": 64, "base_model.model.language_model.model.layers.27.mlp.down_proj": 4, "base_model.model.language_model.model.layers.28.mlp.gate_proj": 64, "base_model.model.language_model.model.layers.28.mlp.up_proj": 8, "base_model.model.language_model.model.layers.28.mlp.down_proj": 16, "base_model.model.language_model.model.layers.29.mlp.gate_proj": 16, "base_model.model.language_model.model.layers.29.mlp.up_proj": 8, "base_model.model.language_model.model.layers.29.mlp.down_proj": 64, "base_model.model.language_model.model.layers.30.mlp.gate_proj": 4, "base_model.model.language_model.model.layers.30.mlp.up_proj": 64, "base_model.model.language_model.model.layers.30.mlp.down_proj": 32, "base_model.model.language_model.model.layers.31.mlp.gate_proj": 4, "base_model.model.language_model.model.layers.31.mlp.up_proj": 32, "base_model.model.language_model.model.layers.31.mlp.down_proj": 32, "base_model.model.language_model.model.layers.32.mlp.gate_proj": 8, "base_model.model.language_model.model.layers.32.mlp.up_proj": 4, "base_model.model.language_model.model.layers.32.mlp.down_proj": 64, "base_model.model.language_model.model.layers.33.mlp.gate_proj": 32, "base_model.model.language_model.model.layers.33.mlp.up_proj": 8, "base_model.model.language_model.model.layers.33.mlp.down_proj": 4, "base_model.model.language_model.model.layers.34.mlp.gate_proj": 64, "base_model.model.language_model.model.layers.34.mlp.up_proj": 16, "base_model.model.language_model.model.layers.34.mlp.down_proj": 64, "base_model.model.language_model.model.layers.35.mlp.gate_proj": 64, "base_model.model.language_model.model.layers.35.mlp.up_proj": 16, "base_model.model.language_model.model.layers.35.mlp.down_proj": 4, "base_model.model.language_model.model.layers.36.mlp.gate_proj": 8, "base_model.model.language_model.model.layers.36.mlp.up_proj": 64, "base_model.model.language_model.model.layers.36.mlp.down_proj": 64, "base_model.model.language_model.model.layers.37.mlp.gate_proj": 32, "base_model.model.language_model.model.layers.37.mlp.up_proj": 4, "base_model.model.language_model.model.layers.37.mlp.down_proj": 64, "base_model.model.language_model.model.layers.38.mlp.gate_proj": 16, "base_model.model.language_model.model.layers.38.mlp.up_proj": 32, "base_model.model.language_model.model.layers.38.mlp.down_proj": 4, "base_model.model.language_model.model.layers.39.mlp.gate_proj": 4, "base_model.model.language_model.model.layers.39.mlp.up_proj": 64, "base_model.model.language_model.model.layers.39.mlp.down_proj": 8}
|