π Model Description
Quantization made by Richard Erkhov.
codegemma-7b-it - GGUF
- Model creator: https://huggingface.co/unsloth/
- Original model: https://huggingface.co/unsloth/codegemma-7b-it/
| Name | Quant method | Size |
|---|---|---|
| codegemma-7b-it.Q2K.gguf | Q2K | 3.24GB |
| codegemma-7b-it.IQ3XS.gguf | IQ3XS | 3.54GB |
| codegemma-7b-it.IQ3S.gguf | IQ3S | 3.71GB |
| codegemma-7b-it.Q3KS.gguf | Q3K_S | 3.71GB |
| codegemma-7b-it.IQ3M.gguf | IQ3M | 3.82GB |
| codegemma-7b-it.Q3K.gguf | Q3K | 4.07GB |
| codegemma-7b-it.Q3KM.gguf | Q3K_M | 4.07GB |
| codegemma-7b-it.Q3KL.gguf | Q3K_L | 4.39GB |
| codegemma-7b-it.IQ4XS.gguf | IQ4XS | 4.48GB |
| codegemma-7b-it.Q40.gguf | Q40 | 4.67GB |
| codegemma-7b-it.IQ4NL.gguf | IQ4NL | 4.69GB |
| codegemma-7b-it.Q4KS.gguf | Q4K_S | 4.7GB |
| codegemma-7b-it.Q4K.gguf | Q4K | 4.96GB |
| codegemma-7b-it.Q4KM.gguf | Q4K_M | 4.96GB |
| codegemma-7b-it.Q41.gguf | Q41 | 5.12GB |
| codegemma-7b-it.Q50.gguf | Q50 | 5.57GB |
| codegemma-7b-it.Q5KS.gguf | Q5K_S | 5.57GB |
| codegemma-7b-it.Q5K.gguf | Q5K | 5.72GB |
| codegemma-7b-it.Q5KM.gguf | Q5K_M | 5.72GB |
| codegemma-7b-it.Q51.gguf | Q51 | 6.02GB |
| codegemma-7b-it.Q6K.gguf | Q6K | 6.53GB |
| codegemma-7b-it.Q80.gguf | Q80 | 8.45GB |
Original model description:
language:
- en
library_name: transformers
license: apache-2.0
tags:
- unsloth
- transformers
- gemma
- bnb
Finetune Mistral, Gemma, Llama 2-5x faster with 70% less memory via Unsloth!
We have a Google Colab Tesla T4 notebook for CodeGemma 7b here: https://colab.research.google.com/drive/19lwcRkZQZtX-qzFP3qZBBHZNcMD1hh?usp=sharing
β¨ Finetune for Free
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.
| Unsloth supports | Free Notebooks | Performance | Memory use |
|---|---|---|---|
| Gemma 7b | βΆοΈ Start on Colab | 2.4x faster | 58% less |
| Mistral 7b | βΆοΈ Start on Colab | 2.2x faster | 62% less |
| Llama-2 7b | βΆοΈ Start on Colab | 2.2x faster | 43% less |
| TinyLlama | βΆοΈ Start on Colab | 3.9x faster | 74% less |
| CodeLlama 34b A100 | βΆοΈ Start on Colab | 1.9x faster | 27% less |
| Mistral 7b 1xT4 | βΆοΈ Start on Kaggle | 5x faster\* | 62% less |
| DPO - Zephyr | βΆοΈ Start on Colab | 1.9x faster | 19% less |
- This conversational notebook is useful for ShareGPT ChatML / Vicuna templates.
- This text completion notebook is for raw text. This DPO notebook replicates Zephyr.
- \* Kaggle has 2x T4s, but we use 1. Due to overhead, 1x T4 is 5x faster.
π GGUF File List
| π Filename | π¦ Size | β‘ Download |
|---|---|---|
|
codegemma-7b-it.IQ3_M.gguf
LFS
Q3
|
3.82 GB | Download |
|
codegemma-7b-it.IQ3_S.gguf
LFS
Q3
|
3.71 GB | Download |
|
codegemma-7b-it.IQ3_XS.gguf
LFS
Q3
|
3.54 GB | Download |
|
codegemma-7b-it.IQ4_NL.gguf
LFS
Q4
|
4.69 GB | Download |
|
codegemma-7b-it.IQ4_XS.gguf
LFS
Q4
|
4.48 GB | Download |
|
codegemma-7b-it.Q2_K.gguf
LFS
Q2
|
3.24 GB | Download |
|
codegemma-7b-it.Q3_K.gguf
LFS
Q3
|
4.07 GB | Download |
|
codegemma-7b-it.Q3_K_L.gguf
LFS
Q3
|
4.39 GB | Download |
|
codegemma-7b-it.Q3_K_M.gguf
LFS
Q3
|
4.07 GB | Download |
|
codegemma-7b-it.Q3_K_S.gguf
LFS
Q3
|
3.71 GB | Download |
|
codegemma-7b-it.Q4_0.gguf
Recommended
LFS
Q4
|
4.67 GB | Download |
|
codegemma-7b-it.Q4_1.gguf
LFS
Q4
|
5.12 GB | Download |
|
codegemma-7b-it.Q4_K.gguf
LFS
Q4
|
4.96 GB | Download |
|
codegemma-7b-it.Q4_K_M.gguf
LFS
Q4
|
4.96 GB | Download |
|
codegemma-7b-it.Q4_K_S.gguf
LFS
Q4
|
4.7 GB | Download |
|
codegemma-7b-it.Q5_0.gguf
LFS
Q5
|
5.57 GB | Download |
|
codegemma-7b-it.Q5_1.gguf
LFS
Q5
|
6.02 GB | Download |
|
codegemma-7b-it.Q5_K.gguf
LFS
Q5
|
5.72 GB | Download |
|
codegemma-7b-it.Q5_K_M.gguf
LFS
Q5
|
5.72 GB | Download |
|
codegemma-7b-it.Q5_K_S.gguf
LFS
Q5
|
5.57 GB | Download |
|
codegemma-7b-it.Q6_K.gguf
LFS
Q6
|
6.53 GB | Download |
|
codegemma-7b-it.Q8_0.gguf
LFS
Q8
|
8.45 GB | Download |