File size: 12,840 Bytes
11f86d0 e213c97 11f86d0 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 |
---
quantized_by: bartowski
pipeline_tag: text-generation
datasets:
- allenai/tulu-3-sft-mixture
base_model: allenai/Llama-3.1-Tulu-3-70B-SFT
license: llama3.1
language:
- en
---
## Llamacpp imatrix Quantizations of Llama-3.1-Tulu-3-70B-SFT
Using <a href="https://github.com/ggerganov/llama.cpp/">llama.cpp</a> release <a href="https://github.com/ggerganov/llama.cpp/releases/tag/b4132">b4132</a> for quantization.
Original model: https://huggingface.co/allenai/Llama-3.1-Tulu-3-70B-SFT
All quants made using imatrix option with dataset from [here](https://gist.github.com/bartowski1182/eb213dccb3571f863da82e99418f81e8)
Run them in [LM Studio](https://lmstudio.ai/)
## Prompt format
```
<|system|>
{system_prompt}
<|user|>
{prompt}
<|assistant|>
```
## Download a file (not the whole branch) from below:
| Filename | Quant type | File Size | Split | Description |
| -------- | ---------- | --------- | ----- | ----------- |
| [Llama-3.1-Tulu-3-70B-SFT-Q8_0.gguf](https://huggingface.co/bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF/tree/main/Llama-3.1-Tulu-3-70B-SFT-Q8_0) | Q8_0 | 74.98GB | true | Extremely high quality, generally unneeded but max available quant. |
| [Llama-3.1-Tulu-3-70B-SFT-Q6_K.gguf](https://huggingface.co/bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF/tree/main/Llama-3.1-Tulu-3-70B-SFT-Q6_K) | Q6_K | 57.89GB | true | Very high quality, near perfect, *recommended*. |
| [Llama-3.1-Tulu-3-70B-SFT-Q5_K_M.gguf](https://huggingface.co/bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF/tree/main/Llama-3.1-Tulu-3-70B-SFT-Q5_K_M) | Q5_K_M | 49.95GB | true | High quality, *recommended*. |
| [Llama-3.1-Tulu-3-70B-SFT-Q5_K_S.gguf](https://huggingface.co/bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF/blob/main/Llama-3.1-Tulu-3-70B-SFT-Q5_K_S.gguf) | Q5_K_S | 48.66GB | false | High quality, *recommended*. |
| [Llama-3.1-Tulu-3-70B-SFT-Q4_K_M.gguf](https://huggingface.co/bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF/blob/main/Llama-3.1-Tulu-3-70B-SFT-Q4_K_M.gguf) | Q4_K_M | 42.52GB | false | Good quality, default size for most use cases, *recommended*. |
| [Llama-3.1-Tulu-3-70B-SFT-Q4_K_S.gguf](https://huggingface.co/bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF/blob/main/Llama-3.1-Tulu-3-70B-SFT-Q4_K_S.gguf) | Q4_K_S | 40.35GB | false | Slightly lower quality with more space savings, *recommended*. |
| [Llama-3.1-Tulu-3-70B-SFT-Q4_0.gguf](https://huggingface.co/bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF/blob/main/Llama-3.1-Tulu-3-70B-SFT-Q4_0.gguf) | Q4_0 | 40.12GB | false | Legacy format, generally not worth using over similarly sized formats |
| [Llama-3.1-Tulu-3-70B-SFT-Q4_0_8_8.gguf](https://huggingface.co/bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF/blob/main/Llama-3.1-Tulu-3-70B-SFT-Q4_0_8_8.gguf) | Q4_0_8_8 | 39.97GB | false | Optimized for ARM and AVX inference. Requires 'sve' support for ARM (see details below). *Don't use on Mac*. |
| [Llama-3.1-Tulu-3-70B-SFT-Q3_K_XL.gguf](https://huggingface.co/bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF/blob/main/Llama-3.1-Tulu-3-70B-SFT-Q3_K_XL.gguf) | Q3_K_XL | 38.06GB | false | Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability. |
| [Llama-3.1-Tulu-3-70B-SFT-IQ4_XS.gguf](https://huggingface.co/bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF/blob/main/Llama-3.1-Tulu-3-70B-SFT-IQ4_XS.gguf) | IQ4_XS | 37.90GB | false | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. |
| [Llama-3.1-Tulu-3-70B-SFT-Q3_K_L.gguf](https://huggingface.co/bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF/blob/main/Llama-3.1-Tulu-3-70B-SFT-Q3_K_L.gguf) | Q3_K_L | 37.14GB | false | Lower quality but usable, good for low RAM availability. |
| [Llama-3.1-Tulu-3-70B-SFT-Q3_K_M.gguf](https://huggingface.co/bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF/blob/main/Llama-3.1-Tulu-3-70B-SFT-Q3_K_M.gguf) | Q3_K_M | 34.27GB | false | Low quality. |
| [Llama-3.1-Tulu-3-70B-SFT-IQ3_M.gguf](https://huggingface.co/bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF/blob/main/Llama-3.1-Tulu-3-70B-SFT-IQ3_M.gguf) | IQ3_M | 31.94GB | false | Medium-low quality, new method with decent performance comparable to Q3_K_M. |
| [Llama-3.1-Tulu-3-70B-SFT-Q3_K_S.gguf](https://huggingface.co/bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF/blob/main/Llama-3.1-Tulu-3-70B-SFT-Q3_K_S.gguf) | Q3_K_S | 30.91GB | false | Low quality, not recommended. |
| [Llama-3.1-Tulu-3-70B-SFT-IQ3_XXS.gguf](https://huggingface.co/bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF/blob/main/Llama-3.1-Tulu-3-70B-SFT-IQ3_XXS.gguf) | IQ3_XXS | 27.47GB | false | Lower quality, new method with decent performance, comparable to Q3 quants. |
| [Llama-3.1-Tulu-3-70B-SFT-Q2_K_L.gguf](https://huggingface.co/bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF/blob/main/Llama-3.1-Tulu-3-70B-SFT-Q2_K_L.gguf) | Q2_K_L | 27.40GB | false | Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable. |
| [Llama-3.1-Tulu-3-70B-SFT-Q2_K.gguf](https://huggingface.co/bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF/blob/main/Llama-3.1-Tulu-3-70B-SFT-Q2_K.gguf) | Q2_K | 26.38GB | false | Very low quality but surprisingly usable. |
| [Llama-3.1-Tulu-3-70B-SFT-IQ2_M.gguf](https://huggingface.co/bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF/blob/main/Llama-3.1-Tulu-3-70B-SFT-IQ2_M.gguf) | IQ2_M | 24.12GB | false | Relatively low quality, uses SOTA techniques to be surprisingly usable. |
| [Llama-3.1-Tulu-3-70B-SFT-IQ2_XS.gguf](https://huggingface.co/bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF/blob/main/Llama-3.1-Tulu-3-70B-SFT-IQ2_XS.gguf) | IQ2_XS | 21.14GB | false | Low quality, uses SOTA techniques to be usable. |
| [Llama-3.1-Tulu-3-70B-SFT-IQ2_XXS.gguf](https://huggingface.co/bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF/blob/main/Llama-3.1-Tulu-3-70B-SFT-IQ2_XXS.gguf) | IQ2_XXS | 19.10GB | false | Very low quality, uses SOTA techniques to be usable. |
| [Llama-3.1-Tulu-3-70B-SFT-IQ1_M.gguf](https://huggingface.co/bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF/blob/main/Llama-3.1-Tulu-3-70B-SFT-IQ1_M.gguf) | IQ1_M | 16.75GB | false | Extremely low quality, *not* recommended. |
## Embed/output weights
Some of these quants (Q3_K_XL, Q4_K_L etc) are the standard quantization method with the embeddings and output weights quantized to Q8_0 instead of what they would normally default to.
## Downloading using huggingface-cli
<details>
<summary>Click to view download instructions</summary>
First, make sure you have hugginface-cli installed:
```
pip install -U "huggingface_hub[cli]"
```
Then, you can target the specific file you want:
```
huggingface-cli download bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF --include "Llama-3.1-Tulu-3-70B-SFT-Q4_K_M.gguf" --local-dir ./
```
If the model is bigger than 50GB, it will have been split into multiple files. In order to download them all to a local folder, run:
```
huggingface-cli download bartowski/Llama-3.1-Tulu-3-70B-SFT-GGUF --include "Llama-3.1-Tulu-3-70B-SFT-Q8_0/*" --local-dir ./
```
You can either specify a new local-dir (Llama-3.1-Tulu-3-70B-SFT-Q8_0) or download them all in place (./)
</details>
## Q4_0_X_X information
<details>
<summary>Click to view Q4_0_X_X information</summary>
These are *NOT* for Metal (Apple) or GPU (nvidia/AMD/intel) offloading, only ARM chips (and certain AVX2/AVX512 CPUs).
If you're using an ARM chip, the Q4_0_X_X quants will have a substantial speedup. Check out Q4_0_4_4 speed comparisons [on the original pull request](https://github.com/ggerganov/llama.cpp/pull/5780#pullrequestreview-21657544660)
To check which one would work best for your ARM chip, you can check [AArch64 SoC features](https://gpages.juszkiewicz.com.pl/arm-socs-table/arm-socs.html) (thanks EloyOn!).
If you're using a CPU that supports AVX2 or AVX512 (typically server CPUs and AMD's latest Zen5 CPUs) and are not offloading to a GPU, the Q4_0_8_8 may offer a nice speed as well:
<details>
<summary>Click to view benchmarks on an AVX2 system (EPYC7702)</summary>
| model | size | params | backend | threads | test | t/s | % (vs Q4_0) |
| ------------------------------ | ---------: | ---------: | ---------- | ------: | ------------: | -------------------: |-------------: |
| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | pp512 | 204.03 ± 1.03 | 100% |
| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | pp1024 | 282.92 ± 0.19 | 100% |
| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | pp2048 | 259.49 ± 0.44 | 100% |
| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | tg128 | 39.12 ± 0.27 | 100% |
| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | tg256 | 39.31 ± 0.69 | 100% |
| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | tg512 | 40.52 ± 0.03 | 100% |
| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | pp512 | 301.02 ± 1.74 | 147% |
| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | pp1024 | 287.23 ± 0.20 | 101% |
| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | pp2048 | 262.77 ± 1.81 | 101% |
| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | tg128 | 18.80 ± 0.99 | 48% |
| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | tg256 | 24.46 ± 3.04 | 83% |
| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | tg512 | 36.32 ± 3.59 | 90% |
| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | pp512 | 271.71 ± 3.53 | 133% |
| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | pp1024 | 279.86 ± 45.63 | 100% |
| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | pp2048 | 320.77 ± 5.00 | 124% |
| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | tg128 | 43.51 ± 0.05 | 111% |
| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | tg256 | 43.35 ± 0.09 | 110% |
| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | tg512 | 42.60 ± 0.31 | 105% |
Q4_0_8_8 offers a nice bump to prompt processing and a small bump to text generation
</details>
</details>
## Which file should I choose?
<details>
<summary>Click here for details</summary>
A great write up with charts showing various performances is provided by Artefact2 [here](https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9)
The first thing to figure out is how big a model you can run. To do this, you'll need to figure out how much RAM and/or VRAM you have.
If you want your model running as FAST as possible, you'll want to fit the whole thing on your GPU's VRAM. Aim for a quant with a file size 1-2GB smaller than your GPU's total VRAM.
If you want the absolute maximum quality, add both your system RAM and your GPU's VRAM together, then similarly grab a quant with a file size 1-2GB Smaller than that total.
Next, you'll need to decide if you want to use an 'I-quant' or a 'K-quant'.
If you don't want to think too much, grab one of the K-quants. These are in format 'QX_K_X', like Q5_K_M.
If you want to get more into the weeds, you can check out this extremely useful feature chart:
[llama.cpp feature matrix](https://github.com/ggerganov/llama.cpp/wiki/Feature-matrix)
But basically, if you're aiming for below Q4, and you're running cuBLAS (Nvidia) or rocBLAS (AMD), you should look towards the I-quants. These are in format IQX_X, like IQ3_M. These are newer and offer better performance for their size.
These I-quants can also be used on CPU and Apple Metal, but will be slower than their K-quant equivalent, so speed vs performance is a tradeoff you'll have to decide.
The I-quants are *not* compatible with Vulcan, which is also AMD, so if you have an AMD card double check if you're using the rocBLAS build or the Vulcan build. At the time of writing this, LM Studio has a preview with ROCm support, and other inference engines have specific builds for ROCm.
</details>
## Credits
Thank you kalomaze and Dampf for assistance in creating the imatrix calibration dataset.
Thank you ZeroWw for the inspiration to experiment with embed/output.
Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski
|