Model save
Browse files
README.md
CHANGED
|
@@ -1,12 +1,12 @@
|
|
| 1 |
---
|
| 2 |
license: llama2
|
| 3 |
-
|
| 4 |
tags:
|
| 5 |
- generated_from_trainer
|
| 6 |
-
base_model: codellama/CodeLlama-7b-hf
|
| 7 |
model-index:
|
| 8 |
- name: sql-code-llama
|
| 9 |
results: []
|
|
|
|
| 10 |
---
|
| 11 |
|
| 12 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
@@ -16,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 16 |
|
| 17 |
This model is a fine-tuned version of [codellama/CodeLlama-7b-hf](https://huggingface.co/codellama/CodeLlama-7b-hf) on an unknown dataset.
|
| 18 |
It achieves the following results on the evaluation set:
|
| 19 |
-
- Loss: 0.
|
| 20 |
|
| 21 |
## Model description
|
| 22 |
|
|
@@ -32,6 +32,21 @@ More information needed
|
|
| 32 |
|
| 33 |
## Training procedure
|
| 34 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 35 |
### Training hyperparameters
|
| 36 |
|
| 37 |
The following hyperparameters were used during training:
|
|
@@ -44,23 +59,39 @@ The following hyperparameters were used during training:
|
|
| 44 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 45 |
- lr_scheduler_type: linear
|
| 46 |
- lr_scheduler_warmup_steps: 100
|
| 47 |
-
- training_steps:
|
|
|
|
| 48 |
|
| 49 |
### Training results
|
| 50 |
|
| 51 |
-
| Training Loss | Epoch
|
| 52 |
-
|
| 53 |
-
| 2.
|
| 54 |
-
| 1.
|
| 55 |
-
| 0.
|
| 56 |
-
| 0.
|
| 57 |
-
| 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 58 |
|
| 59 |
|
| 60 |
### Framework versions
|
| 61 |
|
| 62 |
-
- PEFT 0.
|
| 63 |
-
- Transformers 4.
|
| 64 |
-
- Pytorch 2.
|
| 65 |
-
- Datasets 2.
|
| 66 |
-
- Tokenizers 0.
|
|
|
|
| 1 |
---
|
| 2 |
license: llama2
|
| 3 |
+
base_model: codellama/CodeLlama-7b-hf
|
| 4 |
tags:
|
| 5 |
- generated_from_trainer
|
|
|
|
| 6 |
model-index:
|
| 7 |
- name: sql-code-llama
|
| 8 |
results: []
|
| 9 |
+
library_name: peft
|
| 10 |
---
|
| 11 |
|
| 12 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
|
| 16 |
|
| 17 |
This model is a fine-tuned version of [codellama/CodeLlama-7b-hf](https://huggingface.co/codellama/CodeLlama-7b-hf) on an unknown dataset.
|
| 18 |
It achieves the following results on the evaluation set:
|
| 19 |
+
- Loss: 0.4577
|
| 20 |
|
| 21 |
## Model description
|
| 22 |
|
|
|
|
| 32 |
|
| 33 |
## Training procedure
|
| 34 |
|
| 35 |
+
|
| 36 |
+
The following `bitsandbytes` quantization config was used during training:
|
| 37 |
+
- quant_method: bitsandbytes
|
| 38 |
+
- _load_in_8bit: True
|
| 39 |
+
- _load_in_4bit: False
|
| 40 |
+
- llm_int8_threshold: 6.0
|
| 41 |
+
- llm_int8_skip_modules: None
|
| 42 |
+
- llm_int8_enable_fp32_cpu_offload: False
|
| 43 |
+
- llm_int8_has_fp16_weight: False
|
| 44 |
+
- bnb_4bit_quant_type: fp4
|
| 45 |
+
- bnb_4bit_use_double_quant: False
|
| 46 |
+
- bnb_4bit_compute_dtype: float32
|
| 47 |
+
- bnb_4bit_quant_storage: uint8
|
| 48 |
+
- load_in_4bit: False
|
| 49 |
+
- load_in_8bit: True
|
| 50 |
### Training hyperparameters
|
| 51 |
|
| 52 |
The following hyperparameters were used during training:
|
|
|
|
| 59 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 60 |
- lr_scheduler_type: linear
|
| 61 |
- lr_scheduler_warmup_steps: 100
|
| 62 |
+
- training_steps: 400
|
| 63 |
+
- mixed_precision_training: Native AMP
|
| 64 |
|
| 65 |
### Training results
|
| 66 |
|
| 67 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 68 |
+
|:-------------:|:------:|:----:|:---------------:|
|
| 69 |
+
| 2.1953 | 0.0465 | 20 | 2.0335 |
|
| 70 |
+
| 1.1292 | 0.0931 | 40 | 0.8342 |
|
| 71 |
+
| 0.8133 | 0.1396 | 60 | 0.6552 |
|
| 72 |
+
| 0.5873 | 0.1862 | 80 | 0.5861 |
|
| 73 |
+
| 0.4095 | 0.2327 | 100 | 0.5589 |
|
| 74 |
+
| 0.5731 | 0.2792 | 120 | 0.5159 |
|
| 75 |
+
| 0.4221 | 0.3258 | 140 | 0.5039 |
|
| 76 |
+
| 0.6365 | 0.3723 | 160 | 0.5159 |
|
| 77 |
+
| 0.4779 | 0.4188 | 180 | 0.4867 |
|
| 78 |
+
| 0.3584 | 0.4654 | 200 | 0.5007 |
|
| 79 |
+
| 0.5325 | 0.5119 | 220 | 0.4802 |
|
| 80 |
+
| 0.3998 | 0.5585 | 240 | 0.4767 |
|
| 81 |
+
| 0.5952 | 0.6050 | 260 | 0.4777 |
|
| 82 |
+
| 0.4649 | 0.6515 | 280 | 0.4671 |
|
| 83 |
+
| 0.3394 | 0.6981 | 300 | 0.4752 |
|
| 84 |
+
| 0.5084 | 0.7446 | 320 | 0.4669 |
|
| 85 |
+
| 0.3934 | 0.7912 | 340 | 0.4613 |
|
| 86 |
+
| 0.5762 | 0.8377 | 360 | 0.4617 |
|
| 87 |
+
| 0.4563 | 0.8842 | 380 | 0.4586 |
|
| 88 |
+
| 0.345 | 0.9308 | 400 | 0.4577 |
|
| 89 |
|
| 90 |
|
| 91 |
### Framework versions
|
| 92 |
|
| 93 |
+
- PEFT 0.6.0.dev0
|
| 94 |
+
- Transformers 4.44.0.dev0
|
| 95 |
+
- Pytorch 2.2.2+cu121
|
| 96 |
+
- Datasets 2.19.1
|
| 97 |
+
- Tokenizers 0.19.1
|