File size: 4,109 Bytes
bb8ed7b 841898f bb8ed7b 4df9aac bb8ed7b 4df9aac bb8ed7b 3afcbf1 bb8ed7b 3afcbf1 bb8ed7b 32b4745 3afcbf1 bb8ed7b 32b4745 673c9a2 bb8ed7b f9597a8 bb8ed7b |
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 |
---
base_model: meta-llama/Llama-2-70b-hf
datasets:
- HuggingFaceH4/ultrafeedback_binarized
- allenai/tulu-v2-sft-mixture
exported_from: allenai/tulu-2-dpo-70b
language:
- en
library_name: transformers
license: other
license_link: https://allenai.org/impact-license
license_name: ai2-impact-license-low-risk
quantized_by: mradermacher
---
## About
weighted/imatrix quants of https://huggingface.co/allenai/tulu-2-dpo-70b
<!-- provided-files -->
## Usage
If you are unsure how to use GGUF files, refer to one of [TheBloke's
READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for
more details, including on how to concatenate multi-part files.
## Provided Quants
(sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)
| Link | Type | Size/GB | Notes |
|:-----|:-----|--------:|:------|
| [GGUF](https://huggingface.co/mradermacher/tulu-2-dpo-70b-i1-GGUF/resolve/main/tulu-2-dpo-70b.i1-IQ1_S.gguf) | i1-IQ1_S | 15.0 | for the desperate |
| [GGUF](https://huggingface.co/mradermacher/tulu-2-dpo-70b-i1-GGUF/resolve/main/tulu-2-dpo-70b.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 18.7 | |
| [GGUF](https://huggingface.co/mradermacher/tulu-2-dpo-70b-i1-GGUF/resolve/main/tulu-2-dpo-70b.i1-IQ2_XS.gguf) | i1-IQ2_XS | 20.8 | |
| [GGUF](https://huggingface.co/mradermacher/tulu-2-dpo-70b-i1-GGUF/resolve/main/tulu-2-dpo-70b.i1-IQ2_S.gguf) | i1-IQ2_S | 21.8 | |
| [GGUF](https://huggingface.co/mradermacher/tulu-2-dpo-70b-i1-GGUF/resolve/main/tulu-2-dpo-70b.i1-IQ2_M.gguf) | i1-IQ2_M | 23.7 | |
| [GGUF](https://huggingface.co/mradermacher/tulu-2-dpo-70b-i1-GGUF/resolve/main/tulu-2-dpo-70b.i1-Q2_K.gguf) | i1-Q2_K | 25.9 | IQ3_XXS probably better |
| [GGUF](https://huggingface.co/mradermacher/tulu-2-dpo-70b-i1-GGUF/resolve/main/tulu-2-dpo-70b.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 27.0 | lower quality |
| [GGUF](https://huggingface.co/mradermacher/tulu-2-dpo-70b-i1-GGUF/resolve/main/tulu-2-dpo-70b.i1-IQ3_XS.gguf) | i1-IQ3_XS | 28.6 | |
| [GGUF](https://huggingface.co/mradermacher/tulu-2-dpo-70b-i1-GGUF/resolve/main/tulu-2-dpo-70b.i1-IQ3_S.gguf) | i1-IQ3_S | 30.3 | beats Q3_K* |
| [GGUF](https://huggingface.co/mradermacher/tulu-2-dpo-70b-i1-GGUF/resolve/main/tulu-2-dpo-70b.i1-Q3_K_S.gguf) | i1-Q3_K_S | 30.3 | IQ3_XS probably better |
| [GGUF](https://huggingface.co/mradermacher/tulu-2-dpo-70b-i1-GGUF/resolve/main/tulu-2-dpo-70b.i1-IQ3_M.gguf) | i1-IQ3_M | 31.4 | |
| [GGUF](https://huggingface.co/mradermacher/tulu-2-dpo-70b-i1-GGUF/resolve/main/tulu-2-dpo-70b.i1-Q3_K_M.gguf) | i1-Q3_K_M | 33.7 | IQ3_S probably better |
| [GGUF](https://huggingface.co/mradermacher/tulu-2-dpo-70b-i1-GGUF/resolve/main/tulu-2-dpo-70b.i1-Q3_K_L.gguf) | i1-Q3_K_L | 36.6 | IQ3_M probably better |
| [GGUF](https://huggingface.co/mradermacher/tulu-2-dpo-70b-i1-GGUF/resolve/main/tulu-2-dpo-70b.i1-Q4_K_S.gguf) | i1-Q4_K_S | 39.7 | optimal size/speed/quality |
| [GGUF](https://huggingface.co/mradermacher/tulu-2-dpo-70b-i1-GGUF/resolve/main/tulu-2-dpo-70b.i1-Q4_K_M.gguf) | i1-Q4_K_M | 41.8 | fast, recommended |
| [GGUF](https://huggingface.co/mradermacher/tulu-2-dpo-70b-i1-GGUF/resolve/main/tulu-2-dpo-70b.i1-Q5_K_S.gguf) | i1-Q5_K_S | 47.9 | |
| [GGUF](https://huggingface.co/mradermacher/tulu-2-dpo-70b-i1-GGUF/resolve/main/tulu-2-dpo-70b.i1-Q5_K_M.gguf) | i1-Q5_K_M | 49.2 | |
| [PART 1](https://huggingface.co/mradermacher/tulu-2-dpo-70b-i1-GGUF/resolve/main/tulu-2-dpo-70b.i1-Q6_K.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/tulu-2-dpo-70b-i1-GGUF/resolve/main/tulu-2-dpo-70b.i1-Q6_K.gguf.part2of2) | i1-Q6_K | 57.0 | practically like static Q6_K |
Here is a handy graph by ikawrakow comparing some lower-quality quant
types (lower is better):
And here are Artefact2's thoughts on the matter:
https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9
## Thanks
I thank my company, [nethype GmbH](https://www.nethype.de/), for letting
me use its servers and providing upgrades to my workstation to enable
this work in my free time.
<!-- end -->
|