LLM EXPLORER
 LLM Explorer: A Curated Large Language Model Directory and Analytics
Dark Theme
  Agents   Benchmarks   LLM  List    LLM Hosting    LLM Leaderboards    Blog    Newsfeed    Advertise

TomGrc FusionNet 34Bx2 MoE V0.1 Full Linear DPO by cloudyu

 »  All LLMs  »  cloudyu  »  TomGrc FusionNet 34Bx2 MoE V0.1 Full Linear DPO   URL Share it on

  106   

TomGrc FusionNet 34Bx2 MoE V0.1 Full Linear DPO is an open-source language model by cloudyu. Features: 60.8b LLM, VRAM: 121.8GB, Context: 195K, License: apache-2.0, MoE, HF Score: 77.5, LLM Explorer Score: 0.12, Arc: 74.1, HellaSwag: 86.7, MMLU: 76.7, TruthfulQA: 71.3, WinoGrande: 83.4, GSM8K: 72.9.

  Conversational   Endpoints compatible   Mixtral   Moe   Region:us   Safetensors   Sharded   Tensorflow   Yi
Model Card on HF 🤗: https://huggingface.co/cloudyu/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO 

TomGrc FusionNet 34Bx2 MoE V0.1 Full Linear DPO Benchmarks

ARC: 74.06 vs 96.7 (so35)-23.4%
HellaSwag: 86.67 vs 95.3 (gpt4)-9.1%
MMLU: 76.69 vs 88.3 (so35)-13.1%
TruthfulQA: 71.32 vs 59 (gpt4)20.9%
WinoGrande: 83.43 vs 87.5 (gpt4)-4.7%
GSM8K: 72.93 vs 96.4 (so35)-24.3%
LLME Score: 0.12296
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
TomGrc FusionNet 34Bx2 MoE V0.1 Full Linear DPO (cloudyu/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO)
🌟 Advertise your project 🚀

TomGrc FusionNet 34Bx2 MoE V0.1 Full Linear DPO Parameters and Internals

Model Type 
Mixture of Experts (MoE)
Additional Notes 
Metrics not tested
Training Details 
Methodology:
Fine-tuned using Direct Preference Optimization
Training Time:
1 hour
Hardware Used:
H100 GPU
Model Architecture:
All-linear-parameter fine-tuned MoE
LLM NameTomGrc FusionNet 34Bx2 MoE V0.1 Full Linear DPO
Repository 🤗https://huggingface.co/cloudyu/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO 
Model Size60.8b
Required VRAM121.8 GB
Updated2026-03-29
Maintainercloudyu
Model Typemixtral
Model Files  4.9 GB: 1-of-25   4.8 GB: 2-of-25   4.9 GB: 3-of-25   4.8 GB: 4-of-25   4.9 GB: 5-of-25   4.8 GB: 6-of-25   4.9 GB: 7-of-25   5.0 GB: 8-of-25   5.0 GB: 9-of-25   5.0 GB: 10-of-25   5.0 GB: 11-of-25   5.0 GB: 12-of-25   5.0 GB: 13-of-25   5.0 GB: 14-of-25   5.0 GB: 15-of-25   5.0 GB: 16-of-25   5.0 GB: 17-of-25   5.0 GB: 18-of-25   5.0 GB: 19-of-25   5.0 GB: 20-of-25   5.0 GB: 21-of-25   5.0 GB: 22-of-25   5.0 GB: 23-of-25   5.0 GB: 24-of-25   2.8 GB: 25-of-25
Model ArchitectureMixtralForCausalLM
Licenseapache-2.0
Context Length200000
Model Max Length200000
Transformers Version4.37.2
Tokenizer ClassLlamaTokenizer
Padding Token<s>
Vocabulary Size64000
Torch Data Typebfloat16

Best Alternatives to TomGrc FusionNet 34Bx2 MoE V0.1 Full Linear DPO

Best Alternatives
Context / RAM
Downloads
Likes
Mixtral 34Bx2 MoE 60B195K / 121.9 GB8683111
Yi 34Bx2 MoE 60B DPO195K / 121.8 GB82253
Bagel Hermes 2x34B195K / 121.9 GB10516
Yi 34Bx2 MoE 200K195K / 121.9 GB82342
Yi 34Bx2 MoE 60B195K / 121.9 GB815564
FusionNet 34Bx2 MoE V0.1195K / 121.2 GB598
... Cloudyu Mixtral 34Bx2 MoE 60B195K / 121.8 GB840
FusionNet 34Bx2 MoE32K / 121.2 GB6739
...DPO TomGrc FusionNet 34Bx2 MoE32K / 121.8 GB1144
Nous Hermes 2 MoE 2x34B4K / 121.9 GB7350
Note: green Score (e.g. "73.2") means that the model is better than cloudyu/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO.

Rank the TomGrc FusionNet 34Bx2 MoE V0.1 Full Linear DPO Capabilities

🆘 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! 🌟

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? 52721 in total.

Email us: info@extractum.io. Our Privacy Policy  | Terms and Conditions  | Suggest an improvement.
Our Social Media →  
Original data from HuggingFace, OpenCompass and various public git repos.
Check out Ag3ntum — our secure, self-hosted AI agent for server management.
Release v20260328a
  Support LLM Explorer