Adapter based on Mistral 3B instruct
Browse files- adapter_config.json +38 -0
- adapters.safetensors +3 -0
adapter_config.json
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"adapter_path": "jumbo_adapters",
|
| 3 |
+
"batch_size": 2,
|
| 4 |
+
"config": null,
|
| 5 |
+
"data": "cleaned_data/",
|
| 6 |
+
"fine_tune_type": "lora",
|
| 7 |
+
"grad_checkpoint": false,
|
| 8 |
+
"iters": 10000,
|
| 9 |
+
"learning_rate": 2e-05,
|
| 10 |
+
"lora_parameters": {
|
| 11 |
+
"keys": [
|
| 12 |
+
"mlp.gate_proj",
|
| 13 |
+
"mlp.down_proj",
|
| 14 |
+
"self_attn.q_proj",
|
| 15 |
+
"mlp.up_proj",
|
| 16 |
+
"self_attn.o_proj",
|
| 17 |
+
"self_attn.v_proj",
|
| 18 |
+
"self_attn.k_proj"
|
| 19 |
+
],
|
| 20 |
+
"rank": 8,
|
| 21 |
+
"alpha": 8,
|
| 22 |
+
"dropout": 0.05,
|
| 23 |
+
"scale": 16.0
|
| 24 |
+
},
|
| 25 |
+
"lr_schedule": null,
|
| 26 |
+
"max_seq_length": 2048,
|
| 27 |
+
"model": "ministral/Ministral-3b-instruct",
|
| 28 |
+
"num_layers": 14,
|
| 29 |
+
"resume_adapter_file": null,
|
| 30 |
+
"save_every": 100,
|
| 31 |
+
"seed": 0,
|
| 32 |
+
"steps_per_eval": 200,
|
| 33 |
+
"steps_per_report": 10,
|
| 34 |
+
"test": true,
|
| 35 |
+
"test_batches": 500,
|
| 36 |
+
"train": true,
|
| 37 |
+
"val_batches": 25
|
| 38 |
+
}
|
adapters.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:004a9e19d8d2ae1a2e8a0309eb34c33938e4488a964eaee886f71d9b049f8ee7
|
| 3 |
+
size 36721232
|