RichardErkhov
commited on
Commit
•
66ee0f3
1
Parent(s):
b9bdb50
uploaded readme
Browse files
README.md
ADDED
@@ -0,0 +1,82 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
Quantization made by Richard Erkhov.
|
2 |
+
|
3 |
+
[Github](https://github.com/RichardErkhov)
|
4 |
+
|
5 |
+
[Discord](https://discord.gg/pvy7H8DZMG)
|
6 |
+
|
7 |
+
[Request more models](https://github.com/RichardErkhov/quant_request)
|
8 |
+
|
9 |
+
|
10 |
+
llama-3-8b - GGUF
|
11 |
+
- Model creator: https://huggingface.co/unsloth/
|
12 |
+
- Original model: https://huggingface.co/unsloth/llama-3-8b/
|
13 |
+
|
14 |
+
|
15 |
+
| Name | Quant method | Size |
|
16 |
+
| ---- | ---- | ---- |
|
17 |
+
| [llama-3-8b.Q2_K.gguf](https://huggingface.co/RichardErkhov/unsloth_-_llama-3-8b-gguf/blob/main/llama-3-8b.Q2_K.gguf) | Q2_K | 2.96GB |
|
18 |
+
| [llama-3-8b.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/unsloth_-_llama-3-8b-gguf/blob/main/llama-3-8b.IQ3_XS.gguf) | IQ3_XS | 3.28GB |
|
19 |
+
| [llama-3-8b.IQ3_S.gguf](https://huggingface.co/RichardErkhov/unsloth_-_llama-3-8b-gguf/blob/main/llama-3-8b.IQ3_S.gguf) | IQ3_S | 3.43GB |
|
20 |
+
| [llama-3-8b.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/unsloth_-_llama-3-8b-gguf/blob/main/llama-3-8b.Q3_K_S.gguf) | Q3_K_S | 3.41GB |
|
21 |
+
| [llama-3-8b.IQ3_M.gguf](https://huggingface.co/RichardErkhov/unsloth_-_llama-3-8b-gguf/blob/main/llama-3-8b.IQ3_M.gguf) | IQ3_M | 3.52GB |
|
22 |
+
| [llama-3-8b.Q3_K.gguf](https://huggingface.co/RichardErkhov/unsloth_-_llama-3-8b-gguf/blob/main/llama-3-8b.Q3_K.gguf) | Q3_K | 3.74GB |
|
23 |
+
| [llama-3-8b.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/unsloth_-_llama-3-8b-gguf/blob/main/llama-3-8b.Q3_K_M.gguf) | Q3_K_M | 3.74GB |
|
24 |
+
| [llama-3-8b.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/unsloth_-_llama-3-8b-gguf/blob/main/llama-3-8b.Q3_K_L.gguf) | Q3_K_L | 4.03GB |
|
25 |
+
| [llama-3-8b.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/unsloth_-_llama-3-8b-gguf/blob/main/llama-3-8b.IQ4_XS.gguf) | IQ4_XS | 4.18GB |
|
26 |
+
| [llama-3-8b.Q4_0.gguf](https://huggingface.co/RichardErkhov/unsloth_-_llama-3-8b-gguf/blob/main/llama-3-8b.Q4_0.gguf) | Q4_0 | 4.34GB |
|
27 |
+
| [llama-3-8b.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/unsloth_-_llama-3-8b-gguf/blob/main/llama-3-8b.IQ4_NL.gguf) | IQ4_NL | 4.38GB |
|
28 |
+
| [llama-3-8b.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/unsloth_-_llama-3-8b-gguf/blob/main/llama-3-8b.Q4_K_S.gguf) | Q4_K_S | 4.37GB |
|
29 |
+
| [llama-3-8b.Q4_K.gguf](https://huggingface.co/RichardErkhov/unsloth_-_llama-3-8b-gguf/blob/main/llama-3-8b.Q4_K.gguf) | Q4_K | 4.58GB |
|
30 |
+
| [llama-3-8b.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/unsloth_-_llama-3-8b-gguf/blob/main/llama-3-8b.Q4_K_M.gguf) | Q4_K_M | 4.58GB |
|
31 |
+
| [llama-3-8b.Q4_1.gguf](https://huggingface.co/RichardErkhov/unsloth_-_llama-3-8b-gguf/blob/main/llama-3-8b.Q4_1.gguf) | Q4_1 | 4.78GB |
|
32 |
+
| [llama-3-8b.Q5_0.gguf](https://huggingface.co/RichardErkhov/unsloth_-_llama-3-8b-gguf/blob/main/llama-3-8b.Q5_0.gguf) | Q5_0 | 5.21GB |
|
33 |
+
| [llama-3-8b.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/unsloth_-_llama-3-8b-gguf/blob/main/llama-3-8b.Q5_K_S.gguf) | Q5_K_S | 5.21GB |
|
34 |
+
| [llama-3-8b.Q5_K.gguf](https://huggingface.co/RichardErkhov/unsloth_-_llama-3-8b-gguf/blob/main/llama-3-8b.Q5_K.gguf) | Q5_K | 5.34GB |
|
35 |
+
| [llama-3-8b.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/unsloth_-_llama-3-8b-gguf/blob/main/llama-3-8b.Q5_K_M.gguf) | Q5_K_M | 5.34GB |
|
36 |
+
| [llama-3-8b.Q5_1.gguf](https://huggingface.co/RichardErkhov/unsloth_-_llama-3-8b-gguf/blob/main/llama-3-8b.Q5_1.gguf) | Q5_1 | 5.65GB |
|
37 |
+
| [llama-3-8b.Q6_K.gguf](https://huggingface.co/RichardErkhov/unsloth_-_llama-3-8b-gguf/blob/main/llama-3-8b.Q6_K.gguf) | Q6_K | 6.14GB |
|
38 |
+
|
39 |
+
|
40 |
+
|
41 |
+
|
42 |
+
Original model description:
|
43 |
+
---
|
44 |
+
language:
|
45 |
+
- en
|
46 |
+
license: llama2
|
47 |
+
library_name: transformers
|
48 |
+
tags:
|
49 |
+
- unsloth
|
50 |
+
- transformers
|
51 |
+
- llama
|
52 |
+
- llama-3
|
53 |
+
---
|
54 |
+
|
55 |
+
# Finetune Mistral, Gemma, Llama 2-5x faster with 70% less memory via Unsloth!
|
56 |
+
|
57 |
+
We have a Google Colab Tesla T4 notebook for Llama-3 8b here: https://colab.research.google.com/drive/135ced7oHytdxu3N2DNe1Z0kqjyYIkDXp?usp=sharing
|
58 |
+
Built with Meta Llama 3
|
59 |
+
|
60 |
+
[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/Discord%20button.png" width="200"/>](https://discord.gg/u54VK8m8tk)
|
61 |
+
[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/buy%20me%20a%20coffee%20button.png" width="200"/>](https://ko-fi.com/unsloth)
|
62 |
+
[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)
|
63 |
+
|
64 |
+
## ✨ Finetune for Free
|
65 |
+
|
66 |
+
All notebooks are **beginner friendly**! Add your dataset, click "Run All", and you'll get a 2x faster finetuned model which can be exported to GGUF, vLLM or uploaded to Hugging Face.
|
67 |
+
|
68 |
+
| Unsloth supports | Free Notebooks | Performance | Memory use |
|
69 |
+
|-----------------|--------------------------------------------------------------------------------------------------------------------------|-------------|----------|
|
70 |
+
| **Llama-3 8b** | [▶️ Start on Colab](https://colab.research.google.com/drive/135ced7oHytdxu3N2DNe1Z0kqjyYIkDXp?usp=sharing) | 2.4x faster | 58% less |
|
71 |
+
| **Gemma 7b** | [▶️ Start on Colab](https://colab.research.google.com/drive/10NbwlsRChbma1v55m8LAPYG15uQv6HLo?usp=sharing) | 2.4x faster | 58% less |
|
72 |
+
| **Mistral 7b** | [▶️ Start on Colab](https://colab.research.google.com/drive/1Dyauq4kTZoLewQ1cApceUQVNcnnNTzg_?usp=sharing) | 2.2x faster | 62% less |
|
73 |
+
| **Llama-2 7b** | [▶️ Start on Colab](https://colab.research.google.com/drive/1lBzz5KeZJKXjvivbYvmGarix9Ao6Wxe5?usp=sharing) | 2.2x faster | 43% less |
|
74 |
+
| **TinyLlama** | [▶️ Start on Colab](https://colab.research.google.com/drive/1AZghoNBQaMDgWJpi4RbffGM1h6raLUj9?usp=sharing) | 3.9x faster | 74% less |
|
75 |
+
| **CodeLlama 34b** A100 | [▶️ Start on Colab](https://colab.research.google.com/drive/1y7A0AxE3y8gdj4AVkl2aZX47Xu3P1wJT?usp=sharing) | 1.9x faster | 27% less |
|
76 |
+
| **Mistral 7b** 1xT4 | [▶️ Start on Kaggle](https://www.kaggle.com/code/danielhanchen/kaggle-mistral-7b-unsloth-notebook) | 5x faster\* | 62% less |
|
77 |
+
| **DPO - Zephyr** | [▶️ Start on Colab](https://colab.research.google.com/drive/15vttTpzzVXv_tJwEk-hIcQ0S9FcEWvwP?usp=sharing) | 1.9x faster | 19% less |
|
78 |
+
|
79 |
+
- This [conversational notebook](https://colab.research.google.com/drive/1Aau3lgPzeZKQ-98h69CCu1UJcvIBLmy2?usp=sharing) is useful for ShareGPT ChatML / Vicuna templates.
|
80 |
+
- This [text completion notebook](https://colab.research.google.com/drive/1ef-tab5bhkvWmBOObepl1WgJvfvSzn5Q?usp=sharing) is for raw text. This [DPO notebook](https://colab.research.google.com/drive/15vttTpzzVXv_tJwEk-hIcQ0S9FcEWvwP?usp=sharing) replicates Zephyr.
|
81 |
+
- \* Kaggle has 2x T4s, but we use 1. Due to overhead, 1x T4 is 5x faster.
|
82 |
+
|