Upload README (1).md
Browse files- README (1).md +98 -0
README (1).md
ADDED
@@ -0,0 +1,98 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: llama2
|
3 |
+
base_model: codellama/CodeLlama-7b-hf
|
4 |
+
tags:
|
5 |
+
- generated_from_trainer
|
6 |
+
model-index:
|
7 |
+
- name: karim_codellama
|
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
|
13 |
+
should probably proofread and complete it, then remove this comment. -->
|
14 |
+
|
15 |
+
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/llm_project/llm_project-org/runs/nb0hywqq)
|
16 |
+
# karim_codellama
|
17 |
+
|
18 |
+
This model is a fine-tuned version of [codellama/CodeLlama-7b-hf](https://huggingface.co/codellama/CodeLlama-7b-hf) on the None dataset.
|
19 |
+
It achieves the following results on the evaluation set:
|
20 |
+
- Loss: 0.1887
|
21 |
+
|
22 |
+
## Model description
|
23 |
+
|
24 |
+
More information needed
|
25 |
+
|
26 |
+
## Intended uses & limitations
|
27 |
+
|
28 |
+
More information needed
|
29 |
+
|
30 |
+
## Training and evaluation data
|
31 |
+
|
32 |
+
More information needed
|
33 |
+
|
34 |
+
## Training procedure
|
35 |
+
|
36 |
+
|
37 |
+
The following `bitsandbytes` quantization config was used during training:
|
38 |
+
- quant_method: bitsandbytes
|
39 |
+
- _load_in_8bit: True
|
40 |
+
- _load_in_4bit: False
|
41 |
+
- llm_int8_threshold: 6.0
|
42 |
+
- llm_int8_skip_modules: None
|
43 |
+
- llm_int8_enable_fp32_cpu_offload: False
|
44 |
+
- llm_int8_has_fp16_weight: False
|
45 |
+
- bnb_4bit_quant_type: fp4
|
46 |
+
- bnb_4bit_use_double_quant: False
|
47 |
+
- bnb_4bit_compute_dtype: float32
|
48 |
+
- bnb_4bit_quant_storage: uint8
|
49 |
+
- load_in_4bit: False
|
50 |
+
- load_in_8bit: True
|
51 |
+
### Training hyperparameters
|
52 |
+
|
53 |
+
The following hyperparameters were used during training:
|
54 |
+
- learning_rate: 0.0003
|
55 |
+
- train_batch_size: 8
|
56 |
+
- eval_batch_size: 8
|
57 |
+
- seed: 42
|
58 |
+
- gradient_accumulation_steps: 4
|
59 |
+
- total_train_batch_size: 32
|
60 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
61 |
+
- lr_scheduler_type: linear
|
62 |
+
- lr_scheduler_warmup_steps: 100
|
63 |
+
- training_steps: 400
|
64 |
+
- mixed_precision_training: Native AMP
|
65 |
+
|
66 |
+
### Training results
|
67 |
+
|
68 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
69 |
+
|:-------------:|:------:|:----:|:---------------:|
|
70 |
+
| 1.146 | 0.0787 | 20 | 1.2504 |
|
71 |
+
| 0.8176 | 0.1573 | 40 | 0.6454 |
|
72 |
+
| 0.6291 | 0.2360 | 60 | 0.4881 |
|
73 |
+
| 0.3068 | 0.3147 | 80 | 0.3589 |
|
74 |
+
| 0.5266 | 0.3933 | 100 | 0.4066 |
|
75 |
+
| 0.302 | 0.4720 | 120 | 0.2728 |
|
76 |
+
| 0.1989 | 0.5506 | 140 | 0.2604 |
|
77 |
+
| 0.3157 | 0.6293 | 160 | 0.2502 |
|
78 |
+
| 0.1768 | 0.7080 | 180 | 0.2285 |
|
79 |
+
| 0.4553 | 0.7866 | 200 | 0.2575 |
|
80 |
+
| 0.2183 | 0.8653 | 220 | 0.2152 |
|
81 |
+
| 0.1815 | 0.9440 | 240 | 0.2148 |
|
82 |
+
| 0.2704 | 1.0226 | 260 | 0.2142 |
|
83 |
+
| 0.1662 | 1.1013 | 280 | 0.2001 |
|
84 |
+
| 0.3306 | 1.1799 | 300 | 0.2065 |
|
85 |
+
| 0.2161 | 1.2586 | 320 | 0.1967 |
|
86 |
+
| 0.1429 | 1.3373 | 340 | 0.1925 |
|
87 |
+
| 0.2892 | 1.4159 | 360 | 0.1927 |
|
88 |
+
| 0.1459 | 1.4946 | 380 | 0.1894 |
|
89 |
+
| 0.3078 | 1.5733 | 400 | 0.1887 |
|
90 |
+
|
91 |
+
|
92 |
+
### Framework versions
|
93 |
+
|
94 |
+
- PEFT 0.6.0.dev0
|
95 |
+
- Transformers 4.41.0.dev0
|
96 |
+
- Pytorch 2.1.2
|
97 |
+
- Datasets 2.19.0
|
98 |
+
- Tokenizers 0.19.1
|