Downloads · 30 days
173
5% of all-time downloads
RichardErkhov/Sao10K_-_L3-70B-Euryale-v2.1-gguf
Sao10K_-_L3-70B-Euryale-v2.1-gguf is a machine learning model from RichardErkhov. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
Downloads · 30 days
173
5% of all-time downloads
All-time downloads
3.6K
Public
Repo size
799 GB
Likes
0
Public
Click a slice to open those files.
.gguf926 GB · 100%
From the Hugging Face model README
Quantization made by Richard Erkhov.
L3-70B-Euryale-v2.1 - GGUF
| Name | Quant method | Size |
|---|---|---|
| L3-70B-Euryale-v2.1.Q2_K.gguf | Q2_K | 24.56GB |
| L3-70B-Euryale-v2.1.IQ3_XS.gguf | IQ3_XS | 27.29GB |
| L3-70B-Euryale-v2.1.IQ3_S.gguf | IQ3_S | 28.79GB |
| L3-70B-Euryale-v2.1.Q3_K_S.gguf | Q3_K_S | 28.79GB |
| L3-70B-Euryale-v2.1.IQ3_M.gguf | IQ3_M | 29.74GB |
| L3-70B-Euryale-v2.1.Q3_K.gguf | Q3_K | 31.91GB |
| L3-70B-Euryale-v2.1.Q3_K_M.gguf | Q3_K_M | 31.91GB |
| L3-70B-Euryale-v2.1.Q3_K_L.gguf | Q3_K_L | 34.59GB |
| L3-70B-Euryale-v2.1.IQ4_XS.gguf | IQ4_XS | 35.64GB |
| L3-70B-Euryale-v2.1.Q4_0.gguf | Q4_0 | 37.22GB |
| L3-70B-Euryale-v2.1.IQ4_NL.gguf | IQ4_NL | 37.58GB |
| L3-70B-Euryale-v2.1.Q4_K_S.gguf | Q4_K_S | 37.58GB |
| L3-70B-Euryale-v2.1.Q4_K.gguf | Q4_K | 39.6GB |
| L3-70B-Euryale-v2.1.Q4_K_M.gguf | Q4_K_M | 39.6GB |
| L3-70B-Euryale-v2.1.Q4_1.gguf | Q4_1 | 41.27GB |
| L3-70B-Euryale-v2.1.Q5_0.gguf | Q5_0 | 45.32GB |
| L3-70B-Euryale-v2.1.Q5_K_S.gguf | Q5_K_S | 45.32GB |
| L3-70B-Euryale-v2.1.Q5_K.gguf | Q5_K | 46.52GB |
| L3-70B-Euryale-v2.1.Q5_K_M.gguf | Q5_K_M | 46.52GB |
| L3-70B-Euryale-v2.1.Q5_1.gguf | Q5_1 | 49.36GB |
| L3-70B-Euryale-v2.1.Q6_K.gguf | Q6_K | 53.91GB |
| L3-70B-Euryale-v2.1.Q8_0.gguf | Q8_0 | 69.83GB |
license: cc-by-nc-4.0 language:

She's back!
Stheno's Sister Model, designed to impress.
- Same Dataset used as Stheno v3.2 -> See notes there.
- LoRA Fine-Tune -> FFT is simply too expensive.
- Trained over 8x H100 SXMs and then some more afterwards.
Testing Notes
- Better prompt adherence.
- Better anatomy / spatial awareness.
- Adapts much better to unique and custom formatting / reply formats.
- Very creative, lots of unique swipes.
- Is not restrictive during roleplays.
- Feels like a big brained version of Stheno.
Likely due to it being a 70B model instead of 8B. Similar vibes comparing back in llama 2, where 70B models were simply much more 'aware' in the subtler areas and contexts a smaller model like a 7B or 13B simply were not able to handle.
Recommended Sampler Settings:
Temperature - 1.17
min_p - 0.075
Repetition Penalty - 1.10
SillyTavern Instruct Settings: <br>Context Template: Llama-3-Instruct-Names <br>Instruct Presets: Euryale-v2.1-Llama-3-Instruct
As per usual, support me here:
Ko-fi: https://ko-fi.com/sao10k
Art by wada_kazu / わだかず (pixiv page private?)