Initial GPTQ model commit
Browse files
README.md
CHANGED
@@ -70,10 +70,10 @@ Each separate quant is in a different branch. See below for instructions on fet
|
|
70 |
| [gptq-4bit-32g-actorder_True](https://huggingface.co/TheBloke/Upstage-Llama-2-70B-instruct-v2-GPTQ/tree/gptq-4bit-32g-actorder_True) | 4 | 32 | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 40.66 GB | Yes | AutoGPTQ | 4-bit, with Act Order and group size 32g. Gives highest possible inference quality, with maximum VRAM usage. Poor AutoGPTQ CUDA speed. |
|
71 |
| [gptq-4bit-64g-actorder_True](https://huggingface.co/TheBloke/Upstage-Llama-2-70B-instruct-v2-GPTQ/tree/gptq-4bit-64g-actorder_True) | 4 | 64 | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 37.99 GB | Yes | AutoGPTQ | 4-bit, with Act Order and group size 64g. Uses less VRAM than 32g, but with slightly lower accuracy. Poor AutoGPTQ CUDA speed. |
|
72 |
| [gptq-4bit-128g-actorder_True](https://huggingface.co/TheBloke/Upstage-Llama-2-70B-instruct-v2-GPTQ/tree/gptq-4bit-128g-actorder_True) | 4 | 128 | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 36.65 GB | Yes | AutoGPTQ | 4-bit, with Act Order and group size 128g. Uses even less VRAM than 64g, but with slightly lower accuracy. Poor AutoGPTQ CUDA speed. |
|
73 |
-
| [gptq-3bit--1g-actorder_True](https://huggingface.co/TheBloke/Upstage-Llama-2-70B-instruct-v2-GPTQ/tree/gptq-3bit--1g-actorder_True) | 3 | None | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) |
|
74 |
-
| [gptq-3bit-128g-actorder_False](https://huggingface.co/TheBloke/Upstage-Llama-2-70B-instruct-v2-GPTQ/tree/gptq-3bit-128g-actorder_False) | 3 | 128 | No | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) |
|
75 |
-
| [gptq-3bit-128g-actorder_True](https://huggingface.co/TheBloke/Upstage-Llama-2-70B-instruct-v2-GPTQ/tree/gptq-3bit-128g-actorder_True) | 3 | 128 | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) |
|
76 |
-
| [gptq-3bit-64g-actorder_True](https://huggingface.co/TheBloke/Upstage-Llama-2-70B-instruct-v2-GPTQ/tree/gptq-3bit-64g-actorder_True) | 3 | 64 | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) |
|
77 |
|
78 |
## How to download from branches
|
79 |
|
|
|
70 |
| [gptq-4bit-32g-actorder_True](https://huggingface.co/TheBloke/Upstage-Llama-2-70B-instruct-v2-GPTQ/tree/gptq-4bit-32g-actorder_True) | 4 | 32 | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 40.66 GB | Yes | AutoGPTQ | 4-bit, with Act Order and group size 32g. Gives highest possible inference quality, with maximum VRAM usage. Poor AutoGPTQ CUDA speed. |
|
71 |
| [gptq-4bit-64g-actorder_True](https://huggingface.co/TheBloke/Upstage-Llama-2-70B-instruct-v2-GPTQ/tree/gptq-4bit-64g-actorder_True) | 4 | 64 | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 37.99 GB | Yes | AutoGPTQ | 4-bit, with Act Order and group size 64g. Uses less VRAM than 32g, but with slightly lower accuracy. Poor AutoGPTQ CUDA speed. |
|
72 |
| [gptq-4bit-128g-actorder_True](https://huggingface.co/TheBloke/Upstage-Llama-2-70B-instruct-v2-GPTQ/tree/gptq-4bit-128g-actorder_True) | 4 | 128 | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 36.65 GB | Yes | AutoGPTQ | 4-bit, with Act Order and group size 128g. Uses even less VRAM than 64g, but with slightly lower accuracy. Poor AutoGPTQ CUDA speed. |
|
73 |
+
| [gptq-3bit--1g-actorder_True](https://huggingface.co/TheBloke/Upstage-Llama-2-70B-instruct-v2-GPTQ/tree/gptq-3bit--1g-actorder_True) | 3 | None | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 26.78 GB | No | AutoGPTQ | 3-bit, with Act Order and no group size. Lowest possible VRAM requirements. May be lower quality than 3-bit 128g. |
|
74 |
+
| [gptq-3bit-128g-actorder_False](https://huggingface.co/TheBloke/Upstage-Llama-2-70B-instruct-v2-GPTQ/tree/gptq-3bit-128g-actorder_False) | 3 | 128 | No | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 28.03 GB | No | AutoGPTQ | 3-bit, with group size 128g but no act-order. Slightly higher VRAM requirements than 3-bit None. |
|
75 |
+
| [gptq-3bit-128g-actorder_True](https://huggingface.co/TheBloke/Upstage-Llama-2-70B-instruct-v2-GPTQ/tree/gptq-3bit-128g-actorder_True) | 3 | 128 | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 28.03 GB | No | AutoGPTQ | 3-bit, with group size 128g and act-order. Higher quality than 128g-False but poor AutoGPTQ CUDA speed. |
|
76 |
+
| [gptq-3bit-64g-actorder_True](https://huggingface.co/TheBloke/Upstage-Llama-2-70B-instruct-v2-GPTQ/tree/gptq-3bit-64g-actorder_True) | 3 | 64 | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 29.30 GB | No | AutoGPTQ | 3-bit, with group size 64g and act-order. Highest quality 3-bit option. Poor AutoGPTQ CUDA speed. |
|
77 |
|
78 |
## How to download from branches
|
79 |
|