Upload 2 files
Browse files- config.json +30 -0
- vocab.txt +0 -0
config.json
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model_type": "llama",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"LlamaForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"vocab_size": 32000,
|
| 7 |
+
"hidden_size": 2048,
|
| 8 |
+
"num_hidden_layers": 24,
|
| 9 |
+
"num_attention_heads": 16,
|
| 10 |
+
"lora_alpha": 16,
|
| 11 |
+
"lora_r": 64,
|
| 12 |
+
"lora_dropout": 0.1,
|
| 13 |
+
"use_cache": true,
|
| 14 |
+
"use_4bit": true,
|
| 15 |
+
"bnb_4bit_compute_dtype": "float16",
|
| 16 |
+
"bnb_4bit_quant_type": "nf4",
|
| 17 |
+
"use_nested_quant": false,
|
| 18 |
+
"fp16": true,
|
| 19 |
+
"bf16": false,
|
| 20 |
+
"per_device_train_batch_size": 16,
|
| 21 |
+
"per_device_eval_batch_size": 16,
|
| 22 |
+
"gradient_accumulation_steps": 1,
|
| 23 |
+
"max_grad_norm": 0.5,
|
| 24 |
+
"learning_rate": 0.0004,
|
| 25 |
+
"weight_decay": 0.0003,
|
| 26 |
+
"optim": "adamw_hf",
|
| 27 |
+
"lr_scheduler_type": "linear",
|
| 28 |
+
"warmup_ratio": 0.1,
|
| 29 |
+
"group_by_length": true
|
| 30 |
+
}
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|