Quantization made by Richard Erkhov.

mistral-dory-12b - GGUF

Model creator: https://huggingface.co/BeaverAI/
Original model: https://huggingface.co/BeaverAI/mistral-dory-12b/

Name	Quant method	Size
mistral-dory-12b.Q2_K.gguf	Q2_K	4.46GB
mistral-dory-12b.IQ3_XS.gguf	IQ3_XS	4.94GB
mistral-dory-12b.IQ3_S.gguf	IQ3_S	5.18GB
mistral-dory-12b.Q3_K_S.gguf	Q3_K_S	5.15GB
mistral-dory-12b.IQ3_M.gguf	IQ3_M	5.33GB
mistral-dory-12b.Q3_K.gguf	Q3_K	5.67GB
mistral-dory-12b.Q3_K_M.gguf	Q3_K_M	5.67GB
mistral-dory-12b.Q3_K_L.gguf	Q3_K_L	6.11GB
mistral-dory-12b.IQ4_XS.gguf	IQ4_XS	6.33GB
mistral-dory-12b.Q4_0.gguf	Q4_0	6.59GB
mistral-dory-12b.IQ4_NL.gguf	IQ4_NL	6.65GB
mistral-dory-12b.Q4_K_S.gguf	Q4_K_S	6.63GB
mistral-dory-12b.Q4_K.gguf	Q4_K	6.96GB
mistral-dory-12b.Q4_K_M.gguf	Q4_K_M	6.96GB
mistral-dory-12b.Q4_1.gguf	Q4_1	7.26GB
mistral-dory-12b.Q5_0.gguf	Q5_0	7.93GB
mistral-dory-12b.Q5_K_S.gguf	Q5_K_S	7.93GB
mistral-dory-12b.Q5_K.gguf	Q5_K	8.13GB
mistral-dory-12b.Q5_K_M.gguf	Q5_K_M	8.13GB
mistral-dory-12b.Q5_1.gguf	Q5_1	8.61GB
mistral-dory-12b.Q6_K.gguf	Q6_K	9.37GB
mistral-dory-12b.Q8_0.gguf	Q8_0	12.13GB

Original model description:

base_model: mistralai/Mistral-Nemo-Base-2407 license: apache-2.0 datasets: - BeaverAI/Nemo-Inst-Tune-ds language: - en library_name: transformers

Dory 12b

redone instruct finetune of mistral nemo 12b. not (E)RP-focused, leave that to drummer.

thanks to twisted for the compute :3

Prompting

alpaca-like:

### System:
[Optional system prompt]

### Instruction:
[Query]

### Response:
[Response]<EOT>

### Instruction:
[...]

Training details

Rank 64 QDoRA, trained on the following data mix:

All of kalomaze/Opus_Instruct_3k
All conversations with a reward model rating above 5 in Magpie-Align/Magpie-Gemma2-Pro-Preview-Filtered
50k of Gryphe/Sonnet3.5-SlimOrcaDedupCleaned
All stories above 4.7 rating and published before 2020 in Fizzarolli/FallingThroughTheSkies-592k-Filtered-Filtered

Additional thanks to @nicoboss for giving me access to his private supercomputer, enabling me to provide many more quants, at much higher speed, than I would otherwise be able to.