File size: 2,543 Bytes
1189ce6
 
632819e
 
 
 
 
1189ce6
 
 
632819e
1189ce6
 
 
 
 
 
 
 
 
 
 
 
 
632819e
1189ce6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
---
base_model: meta-llama/Meta-Llama-3.1-8B-Instruct
datasets:
- barc0/induction_100k-gpt4-description-gpt4omini-code_generated_problems_messages_format_0.3
- barc0/induction_100k_gpt4o-mini_generated_problems_seed100.jsonl_messages_format_0.3
- barc0/induction_gpt-4_description_20000_with_llama_codegen_messages_format_0.3
- barc0/induction_gpt-4_description_20000_with_gpt-4o-mini_codegen_messages_format_0.3
library_name: peft
license: llama3.1
tags:
- alignment-handbook
- trl
- sft
- generated_from_trainer
model-index:
- name: l3.1-8b-inst-lora64-induction-gpt4wmini100k-mini100k-gpt4wmini20k-gpt4wllama20k-lr2e-4-ep3
  results: []
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# l3.1-8b-inst-lora64-induction-gpt4wmini100k-mini100k-gpt4wmini20k-gpt4wllama20k-lr2e-4-ep3

This model is a fine-tuned version of [meta-llama/Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct) on the barc0/induction_100k-gpt4-description-gpt4omini-code_generated_problems_messages_format_0.3, the barc0/induction_100k_gpt4o-mini_generated_problems_seed100.jsonl_messages_format_0.3, the barc0/induction_gpt-4_description_20000_with_llama_codegen_messages_format_0.3 and the barc0/induction_gpt-4_description_20000_with_gpt-4o-mini_codegen_messages_format_0.3 datasets.
It achieves the following results on the evaluation set:
- Loss: 0.2684

## Model description

More information needed

## Intended uses & limitations

More information needed

## Training and evaluation data

More information needed

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 0.0002
- train_batch_size: 8
- eval_batch_size: 4
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 2
- total_train_batch_size: 128
- total_eval_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 3

### Training results

| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 0.2882        | 1.0   | 1784 | 0.2852          |
| 0.257         | 2.0   | 3568 | 0.2705          |
| 0.2329        | 3.0   | 5352 | 0.2684          |


### Framework versions

- PEFT 0.13.0
- Transformers 4.45.0.dev0
- Pytorch 2.4.1+cu121
- Datasets 3.0.0
- Tokenizers 0.19.1