Exllamav2 quant (exl2 / 8.0 bpw) made with ExLlamaV2 v0.1.1

Other EXL2 quants:

Quant	Model Size	lm_head
2.2	7777 MB	6
2.5	8521 MB	6
3.0	9944 MB	6
3.5	11355 MB	6
3.75	12070 MB	6
4.0	12785 MB	6
4.25	13504 MB	6
5.0	15634 MB	6
6.0	18589 MB	8
6.5	19948 MB	8
8.0	24070 MB	8

GGUF

Experimental RP-oriented MoE, the idea was to get a model that would be equal to or better than Mixtral 8x7B and it's finetunes in RP/ERP tasks.

There's:

v1.15A
v1.15B <- You're here

Llama 3 SnowStorm v1.15B 4x8B

base_model: Sao10K_L3-8B-Stheno-v3.1
gate_mode: random
dtype: bfloat16
experts_per_token: 2
experts:
  - source_model: Nitral-AI_Poppy_Porpoise-1.0-L3-8B
  - source_model: NeverSleep_Llama-3-Lumimaid-8B-v0.1-OAS
  - source_model: openlynn_Llama-3-Soliloquy-8B-v2
  - source_model: Sao10K_L3-8B-Stheno-v3.1

Models used

Difference(from SnowStorm v1.0)

Update from ChaoticNeutrals/Poppy_Porpoise-v0.7-L3-8B to Nitral-AI/Poppy_Porpoise-1.0-L3-8B
Change base model from NeverSleep/Llama-3-Lumimaid-8B-v0.1-OAS to Sao10K/L3-8B-Stheno-v3.1

Vision

llama3_mmproj

Prompt format: Llama 3

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	68.01
AI2 Reasoning Challenge (25-Shot)	60.67
HellaSwag (10-Shot)	81.60
MMLU (5-Shot)	68.12
TruthfulQA (0-shot)	51.69
Winogrande (5-shot)	76.56
GSM8k (5-shot)	69.45