RichardErkhov/Ba2han_-_BruinsV2-OpHermesNeu-11B-gguf

Quantization made by Richard Erkhov.

Github

Discord

Request more models

BruinsV2-OpHermesNeu-11B - GGUF

• Model creator: https://huggingface.co/Ba2han/
• Original model: https://huggingface.co/Ba2han/BruinsV2-OpHermesNeu-11B/

Name	Quant method	Size
BruinsV2-OpHermesNeu-11B.Q2_K.gguf	Q2_K	3.73GB
BruinsV2-OpHermesNeu-11B.IQ3_XS.gguf	IQ3_XS	4.14GB
BruinsV2-OpHermesNeu-11B.IQ3_S.gguf	IQ3_S	4.37GB
BruinsV2-OpHermesNeu-11B.Q3_K_S.gguf	Q3_K_S	4.34GB
BruinsV2-OpHermesNeu-11B.IQ3_M.gguf	IQ3_M	4.51GB
BruinsV2-OpHermesNeu-11B.Q3_K.gguf	Q3_K	4.84GB
BruinsV2-OpHermesNeu-11B.Q3_K_M.gguf	Q3_K_M	4.84GB
BruinsV2-OpHermesNeu-11B.Q3_K_L.gguf	Q3_K_L	5.26GB
BruinsV2-OpHermesNeu-11B.IQ4_XS.gguf	IQ4_XS	5.43GB
BruinsV2-OpHermesNeu-11B.Q4_0.gguf	Q4_0	5.66GB
BruinsV2-OpHermesNeu-11B.IQ4_NL.gguf	IQ4_NL	5.72GB
BruinsV2-OpHermesNeu-11B.Q4_K_S.gguf	Q4_K_S	5.7GB
BruinsV2-OpHermesNeu-11B.Q4_K.gguf	Q4_K	6.02GB
BruinsV2-OpHermesNeu-11B.Q4_K_M.gguf	Q4_K_M	6.02GB
BruinsV2-OpHermesNeu-11B.Q4_1.gguf	Q4_1	6.27GB
BruinsV2-OpHermesNeu-11B.Q5_0.gguf	Q5_0	6.89GB
BruinsV2-OpHermesNeu-11B.Q5_K_S.gguf	Q5_K_S	6.89GB
BruinsV2-OpHermesNeu-11B.Q5_K.gguf	Q5_K	7.08GB
BruinsV2-OpHermesNeu-11B.Q5_K_M.gguf	Q5_K_M	7.08GB
BruinsV2-OpHermesNeu-11B.Q5_1.gguf	Q5_1	7.51GB
BruinsV2-OpHermesNeu-11B.Q6_K.gguf	Q6_K	8.2GB
BruinsV2-OpHermesNeu-11B.Q8_0.gguf	Q8_0	10.62GB

Original model description:

license: mit

Task	Version	Metric	Value		Stderr
arc_challenge	0	acc	0.6527	±	0.0139
		acc_norm	0.6869	±	0.0136

Warning! This model may or may not be contaminated See discussion 474. What a shame. It still does perform well though

A passthrough merge of OpenHermes-2.5-neural-chat-7b-v3-1 and Bruins-V2. To be updated.

Template: ChatML

My settings:

Temperature: 0.7-0.8

Min_p: 0.12

Top_K: 0

Repetition Penalty: 1.16

Mirostat Tau: 2.5-3

Mirostat Eta: 0.12

Personal Thoughts:

• The model sometimes throws wrong tags, you can add those to "Custom stopping strings" in Oobabooga.
• Output with Mirostat consistently felt smarter than a set Top_K rate.

Note: The model is hallucinating hard in chat mode for me in some instances, like writing adblocker messages. Kind of funny.

I am not sure which dataset involved was poisoned.