95 lines
5.6 KiB
Markdown
95 lines
5.6 KiB
Markdown
---
|
||
base_model: seedboxai/KafkaLM-15B
|
||
language:
|
||
- en
|
||
- de
|
||
- fr
|
||
- es
|
||
- it
|
||
- pt
|
||
library_name: transformers
|
||
license: apache-2.0
|
||
mradermacher:
|
||
readme_rev: 1
|
||
quantized_by: mradermacher
|
||
tags:
|
||
- pruning
|
||
- distillation
|
||
- sparsity‑2:4
|
||
---
|
||
## About
|
||
|
||
<!-- ### quantize_version: 2 -->
|
||
<!-- ### output_tensor_quantised: 1 -->
|
||
<!-- ### convert_type: hf -->
|
||
<!-- ### vocab_type: -->
|
||
<!-- ### tags: nicoboss -->
|
||
<!-- ### quants: Q2_K IQ3_M Q4_K_S IQ3_XXS Q3_K_M small-IQ4_NL Q4_K_M IQ2_M Q6_K IQ4_XS Q2_K_S IQ1_M Q3_K_S IQ2_XXS Q3_K_L IQ2_XS Q5_K_S IQ2_S IQ1_S Q5_K_M Q4_0 IQ3_XS Q4_1 IQ3_S -->
|
||
<!-- ### quants_skip: -->
|
||
<!-- ### skip_mmproj: -->
|
||
weighted/imatrix quants of https://huggingface.co/seedboxai/KafkaLM-15B
|
||
|
||
<!-- provided-files -->
|
||
|
||
***For a convenient overview and download list, visit our [model page for this model](https://hf.tst.eu/model#KafkaLM-15B-i1-GGUF).***
|
||
|
||
static quants are available at https://huggingface.co/mradermacher/KafkaLM-15B-GGUF
|
||
## 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/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.imatrix.gguf) | imatrix | 0.1 | imatrix file (for creating your own quants) |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-IQ1_S.gguf) | i1-IQ1_S | 3.7 | for the desperate |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-IQ1_M.gguf) | i1-IQ1_M | 4.1 | mostly desperate |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 4.6 | |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-IQ2_XS.gguf) | i1-IQ2_XS | 5.0 | |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-IQ2_S.gguf) | i1-IQ2_S | 5.2 | |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-IQ2_M.gguf) | i1-IQ2_M | 5.6 | |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-Q2_K_S.gguf) | i1-Q2_K_S | 5.7 | very low quality |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-Q2_K.gguf) | i1-Q2_K | 6.1 | IQ3_XXS probably better |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 6.3 | lower quality |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-IQ3_XS.gguf) | i1-IQ3_XS | 6.8 | |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-Q3_K_S.gguf) | i1-Q3_K_S | 7.1 | IQ3_XS probably better |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-IQ3_S.gguf) | i1-IQ3_S | 7.1 | beats Q3_K* |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-IQ3_M.gguf) | i1-IQ3_M | 7.3 | |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-Q3_K_M.gguf) | i1-Q3_K_M | 7.8 | IQ3_S probably better |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-Q3_K_L.gguf) | i1-Q3_K_L | 8.4 | IQ3_M probably better |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-IQ4_XS.gguf) | i1-IQ4_XS | 8.6 | |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-IQ4_NL.gguf) | i1-IQ4_NL | 9.1 | prefer IQ4_XS |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-Q4_0.gguf) | i1-Q4_0 | 9.1 | fast, low quality |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-Q4_K_S.gguf) | i1-Q4_K_S | 9.1 | optimal size/speed/quality |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-Q4_K_M.gguf) | i1-Q4_K_M | 9.6 | fast, recommended |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-Q4_1.gguf) | i1-Q4_1 | 10.0 | |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-Q5_K_S.gguf) | i1-Q5_K_S | 10.9 | |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-Q5_K_M.gguf) | i1-Q5_K_M | 11.2 | |
|
||
| [GGUF](https://huggingface.co/mradermacher/KafkaLM-15B-i1-GGUF/resolve/main/KafkaLM-15B.i1-Q6_K.gguf) | i1-Q6_K | 12.9 | 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
|
||
|
||
## FAQ / Model Request
|
||
|
||
See https://huggingface.co/mradermacher/model_requests for some answers to
|
||
questions you might have and/or if you want some other model quantized.
|
||
|
||
## 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. Additional thanks to [@nicoboss](https://huggingface.co/nicoboss) for giving me access to his private supercomputer, enabling me to provide many more imatrix quants, at much higher quality, than I would otherwise be able to.
|
||
|
||
<!-- end -->
|