Commit
·
bc8723f
1
Parent(s):
e74bc84
End of training
Browse files
README.md
CHANGED
|
@@ -30,6 +30,28 @@ More information needed
|
|
| 30 |
|
| 31 |
## Training procedure
|
| 32 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 33 |
### Training hyperparameters
|
| 34 |
|
| 35 |
The following hyperparameters were used during training:
|
|
@@ -48,8 +70,8 @@ The following hyperparameters were used during training:
|
|
| 48 |
|
| 49 |
### Framework versions
|
| 50 |
|
| 51 |
-
- PEFT 0.7.
|
| 52 |
-
- Transformers 4.36.0
|
| 53 |
- Pytorch 2.1.0+cu118
|
| 54 |
- Datasets 2.15.0
|
| 55 |
- Tokenizers 0.15.0
|
|
|
|
| 30 |
|
| 31 |
## Training procedure
|
| 32 |
|
| 33 |
+
|
| 34 |
+
The following `bitsandbytes` quantization config was used during training:
|
| 35 |
+
- quant_method: gptq
|
| 36 |
+
- bits: 4
|
| 37 |
+
- tokenizer: None
|
| 38 |
+
- dataset: None
|
| 39 |
+
- group_size: 128
|
| 40 |
+
- damp_percent: 0.1
|
| 41 |
+
- desc_act: True
|
| 42 |
+
- sym: True
|
| 43 |
+
- true_sequential: True
|
| 44 |
+
- use_cuda_fp16: False
|
| 45 |
+
- model_seqlen: 4096
|
| 46 |
+
- block_name_to_quantize: model.layers
|
| 47 |
+
- module_name_preceding_first_block: ['model.embed_tokens']
|
| 48 |
+
- batch_size: 1
|
| 49 |
+
- pad_token_id: None
|
| 50 |
+
- use_exllama: False
|
| 51 |
+
- max_input_length: None
|
| 52 |
+
- exllama_config: {'version': <ExllamaVersion.ONE: 1>}
|
| 53 |
+
- cache_block_outputs: True
|
| 54 |
+
|
| 55 |
### Training hyperparameters
|
| 56 |
|
| 57 |
The following hyperparameters were used during training:
|
|
|
|
| 70 |
|
| 71 |
### Framework versions
|
| 72 |
|
| 73 |
+
- PEFT 0.7.0
|
| 74 |
+
- Transformers 4.36.0
|
| 75 |
- Pytorch 2.1.0+cu118
|
| 76 |
- Datasets 2.15.0
|
| 77 |
- Tokenizers 0.15.0
|