File size: 6,199 Bytes
a68226c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
Quantization made by Richard Erkhov.

[Github](https://github.com/RichardErkhov)

[Discord](https://discord.gg/pvy7H8DZMG)

[Request more models](https://github.com/RichardErkhov/quant_request)


llama3-8b-math-sft-subtask-7 - GGUF
- Model creator: https://huggingface.co/Dynosaur/
- Original model: https://huggingface.co/Dynosaur/llama3-8b-math-sft-subtask-7/


| Name | Quant method | Size |
| ---- | ---- | ---- |
| [llama3-8b-math-sft-subtask-7.Q2_K.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.Q2_K.gguf) | Q2_K | 2.96GB |
| [llama3-8b-math-sft-subtask-7.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.IQ3_XS.gguf) | IQ3_XS | 3.28GB |
| [llama3-8b-math-sft-subtask-7.IQ3_S.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.IQ3_S.gguf) | IQ3_S | 3.43GB |
| [llama3-8b-math-sft-subtask-7.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.Q3_K_S.gguf) | Q3_K_S | 3.41GB |
| [llama3-8b-math-sft-subtask-7.IQ3_M.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.IQ3_M.gguf) | IQ3_M | 3.52GB |
| [llama3-8b-math-sft-subtask-7.Q3_K.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.Q3_K.gguf) | Q3_K | 3.74GB |
| [llama3-8b-math-sft-subtask-7.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.Q3_K_M.gguf) | Q3_K_M | 3.74GB |
| [llama3-8b-math-sft-subtask-7.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.Q3_K_L.gguf) | Q3_K_L | 4.03GB |
| [llama3-8b-math-sft-subtask-7.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.IQ4_XS.gguf) | IQ4_XS | 4.18GB |
| [llama3-8b-math-sft-subtask-7.Q4_0.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.Q4_0.gguf) | Q4_0 | 4.34GB |
| [llama3-8b-math-sft-subtask-7.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.IQ4_NL.gguf) | IQ4_NL | 4.38GB |
| [llama3-8b-math-sft-subtask-7.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.Q4_K_S.gguf) | Q4_K_S | 4.37GB |
| [llama3-8b-math-sft-subtask-7.Q4_K.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.Q4_K.gguf) | Q4_K | 4.58GB |
| [llama3-8b-math-sft-subtask-7.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.Q4_K_M.gguf) | Q4_K_M | 4.58GB |
| [llama3-8b-math-sft-subtask-7.Q4_1.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.Q4_1.gguf) | Q4_1 | 4.78GB |
| [llama3-8b-math-sft-subtask-7.Q5_0.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.Q5_0.gguf) | Q5_0 | 5.21GB |
| [llama3-8b-math-sft-subtask-7.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.Q5_K_S.gguf) | Q5_K_S | 5.21GB |
| [llama3-8b-math-sft-subtask-7.Q5_K.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.Q5_K.gguf) | Q5_K | 5.34GB |
| [llama3-8b-math-sft-subtask-7.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.Q5_K_M.gguf) | Q5_K_M | 5.34GB |
| [llama3-8b-math-sft-subtask-7.Q5_1.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.Q5_1.gguf) | Q5_1 | 5.65GB |
| [llama3-8b-math-sft-subtask-7.Q6_K.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.Q6_K.gguf) | Q6_K | 6.14GB |
| [llama3-8b-math-sft-subtask-7.Q8_0.gguf](https://huggingface.co/RichardErkhov/Dynosaur_-_llama3-8b-math-sft-subtask-7-gguf/blob/main/llama3-8b-math-sft-subtask-7.Q8_0.gguf) | Q8_0 | 7.95GB |




Original model description:
---
library_name: transformers
license: llama3
base_model: Dynosaur/llama3-8b-math-sft
tags:
- alignment-handbook
- trl
- sft
- generated_from_trainer
- trl
- sft
- generated_from_trainer
datasets:
- Dynosaur/math-sft-subtask-7
model-index:
- name: llama3-8b-math-sft-subtask-7
  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. -->

# llama3-8b-math-sft-subtask-7

This model is a fine-tuned version of [Dynosaur/llama3-8b-math-sft](https://huggingface.co/Dynosaur/llama3-8b-math-sft) on the Dynosaur/math-sft-subtask-7 dataset.

## 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: 2e-05
- train_batch_size: 2
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 8
- total_train_batch_size: 128
- total_eval_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 2

### Training results



### Framework versions

- Transformers 4.44.2
- Pytorch 2.4.1+cu121
- Datasets 3.0.0
- Tokenizers 0.19.1