Synthia MoE V3 Mixtral 8x7B 3.5bpw H6 EXL2 2 by LoneStriker

 ยป  All LLMs  ยป  LoneStriker  ยป  Synthia MoE V3 Mixtral 8x7B 3.5bpw H6 EXL2 2   URL Share it on

  Autotrain compatible   Endpoints compatible   Exl2   Mixtral   Moe   Pytorch   Quantized   Region:us   Sharded   Tensorflow

Synthia MoE V3 Mixtral 8x7B 3.5bpw H6 EXL2 2 Benchmarks

nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Synthia MoE V3 Mixtral 8x7B 3.5bpw H6 EXL2 2 (LoneStriker/Synthia-MoE-v3-Mixtral-8x7B-3.5bpw-h6-exl2-2)

Synthia MoE V3 Mixtral 8x7B 3.5bpw H6 EXL2 2 Parameters and Internals

Use Cases 
Primary Use Cases:
Answering complex questions with reasoning
Additional Notes 
Model might be overfitted due to higher learning rate, which is expected to be fixed in the next release.
Training Details 
Data Sources:
Synthia-v3.0 dataset
Data Volume:
~10K super high-quality GPT-4-Turbo generated samples
Methodology:
Trained on the Orca-2 principle of replacing the system context with one message. No system context included.
LLM NameSynthia MoE V3 Mixtral 8x7B 3.5bpw H6 EXL2 2
Repository ๐Ÿค—https://huggingface.co/LoneStriker/Synthia-MoE-v3-Mixtral-8x7B-3.5bpw-h6-exl2-2 
Required VRAM20.7 GB
Updated2024-12-21
MaintainerLoneStriker
Model Typemixtral
Model Files  8.6 GB: 1-of-3   8.6 GB: 2-of-3   3.5 GB: 3-of-3
Quantization Typeexl2
Model ArchitectureMixtralForCausalLM
Licenseapache-2.0
Context Length32768
Model Max Length32768
Transformers Version4.36.0.dev0
Tokenizer ClassLlamaTokenizer
Padding Token</s>
Vocabulary Size32000
Torch Data Typebfloat16

Best Alternatives to Synthia MoE V3 Mixtral 8x7B 3.5bpw H6 EXL2 2

Best Alternatives
Context / RAM
Downloads
Likes
...oE V0.1 DPO F16 5.0bpw H6 EXL2195K / 38.8 GB60
...oE V0.1 DPO F16 4.0bpw H6 EXL2195K / 31.3 GB50
WizardLM 2 8x22 EXL2 4.0bpw64K / 70.9 GB121
...2 Mixtral 8x22b 6.0bpw H8 EXL264K / 105.8 GB81
...M 2 8x22B Beige 2.4bpw H6 EXL264K / 42.7 GB150
...M 2 8x22B Beige 3.0bpw H6 EXL264K / 53.2 GB140
...M 2 8x22B Beige 4.0bpw H6 EXL264K / 70.8 GB120
...rdLM 2 8x22B Beige EXL2 5.0bpw64K / 88.4 GB110
...M 2 8x22B Beige 5.0bpw H6 EXL264K / 88.5 GB80
...B Instruct V0.1 8.0bpw H8 EXL264K / 120.2 GB61
Note: green Score (e.g. "73.2") means that the model is better than LoneStriker/Synthia-MoE-v3-Mixtral-8x7B-3.5bpw-h6-exl2-2.

Rank the Synthia MoE V3 Mixtral 8x7B 3.5bpw H6 EXL2 2 Capabilities

๐Ÿ†˜ Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! ๐ŸŒŸ

Instruction Following and Task Automation  
Factuality and Completeness of Knowledge  
Censorship and Alignment  
Data Analysis and Insight Generation  
Text Generation  
Text Summarization and Feature Extraction  
Code Generation  
Multi-Language Support and Translation  

What open-source LLMs or SLMs are you in search of? 40013 in total.

Our Social Media →  
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241217