File size: 2,933 Bytes
b85856b 533b4d2 fb3301f b85856b 533b4d2 5ae74ce 533b4d2 5ae74ce b8a588a 7144ac6 533b4d2 50307b6 4533b02 55b4921 4533b02 533b4d2 50307b6 5ae74ce 50307b6 533b4d2 5ae74ce 533b4d2 |
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 |
---
tags:
- gguf
- iMat
- conversational
- storywriting
license: cc-by-nc-4.0
---
<h3> Model Card for Fimbulvetr-11B-v2-iMat-GGUF</h3>
* Model creator: [Sao10K](https://huggingface.co/Sao10K/)
* Original model: [Fimbulvetr-11B-v2](https://huggingface.co/Sao10K/Fimbulvetr-11B-v2)
<b>Update 3/4/24: </b> Newest I-Quant format <b>[IQ4_XS](https://huggingface.co/InferenceIllusionist/Fimbulvetr-11B-v2-iMat-GGUF/blob/main/Fimbulvetr-11B-v2-iMat-IQ4_XS.gguf)</b> shows superior performance to previous I-quants @ a whopping 4.25 bpw in [benchmarks](https://github.com/ggerganov/llama.cpp/pull/5747)
Tested on latest llama.cpp & koboldcpp v.1.60.
<h4> This model fits a whole lot into its size! Impressed by its understanding of other languages</h4>
<img src="https://huggingface.co/InferenceIllusionist/Fimbulvetr-11B-v2-iMat-GGUF/resolve/main/Fimbulvetr-11B-v2%20IQ4_XS.JPG" width="850"/>
<b>Tip: Select the biggest size that you can fit in VRAM while still allowing some space for context</b>
All credits to Sao10K for the original model. This is just a quick test of the new quantization types such as IQ_3S in an attempt to further reduce VRAM requirements.
Quantized from fp16 with love. Importance matrix file [Fimbulvetr-11B-v2-imatrix.dat](https://huggingface.co/InferenceIllusionist/Fimbulvetr-11B-v2-iMat-GGUF/blob/main/Fimbulvetr-11B-v2-imatrix.dat) was calculated using Q8_0.
<i>Looking for Q3/Q4/Q5 quants? See the link in the original model card below.</i>
---
*Cute girl to catch your attention.*
**https://huggingface.co/Sao10K/Fimbulvetr-11B-v2-GGUF <------ GGUF**
# Fimbulvetr-v2 - A Solar-Based Model
Prompt Formats - Alpaca or Vicuna. Either one works fine.
Recommended SillyTavern Presets - Universal Light
Alpaca:
```
### Instruction:
<Prompt>
### Input:
<Insert Context Here>
### Response:
```
Vicuna:
```
System: <Prompt>
User: <Input>
Assistant:
```
****
Changelogs:
25/2 - repo renamed to remove test, model card redone. Model's officially out.
<br>15/2 - Heavy testing complete. Good feedback.
***
<details><summary>Rant - Kept For Historical Reasons</summary>
Ramble to meet minimum length requirements:
Tbh i wonder if this shit is even worth doing. Like im just some broke guy lmao I've spent so much. And for what? I guess creds. Feels good when a model gets good feedback, but it seems like im invisible sometimes. I should be probably advertising myself and my models on other places but I rarely have the time to. Probably just internal jealousy sparking up here and now. Wahtever I guess.
Anyway cool EMT vocation I'm doing is cool except it pays peanuts, damn bruh 1.1k per month lmao. Government to broke to pay for shit. Pays the bills I suppose.
Anyway cool beans, I'm either going to continue the Solar Train or go to Mixtral / Yi when I get paid.
You still here?
</details><br> |