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

Einstein V4 Qwen 1.5 32B by Weyaxi

 »  All LLMs  »  Weyaxi  »  Einstein V4 Qwen 1.5 32B   URL Share it on

  89   
  Arxiv:2305.14314   Autotrain compatible   Axolotl Base model:finetune:qwen/qwen1...   Base model:qwen/qwen1.5-32b   Biology   Chatml   Chemistry   Conversational   Dataset:allenai/ai2 arc   Dataset:bigbio/med qa   Dataset:camel-ai/biology   Dataset:camel-ai/chemistry   Dataset:camel-ai/math   Dataset:camel-ai/physics Dataset:cot-alpaca-gpt4-from-o...   Dataset:derek-thomas/scienceqa Dataset:glaiveai/glaive-code-a... Dataset:jondurbin/airoboros-3.... Dataset:knowrohit07/saraswati-...   Dataset:ldjnr/capybara   Dataset:lmsys/lmsys-chat-1m   Dataset:mandyyyyii/scibench Dataset:meta-math/metamathqa-4...   Dataset:metaeval/reclor Dataset:migtissera/synthia-v1....   Dataset:open-orca/slimorca   Dataset:openbookqa   Dataset:piqa   Dataset:sablo/oasst2 curated   Dataset:scibench   Dataset:sciq Dataset:stem-ai-mtl/electrical...   Dataset:tiger-lab/mathinstruct   Dataset:tiger-lab/scienceeval   Einstein   En   Endpoints compatible   Finetuned   Generated from trainer   Gpt4   Instruct   Math   Phi   Phi2   Physics   Qwen2   Region:us   Safetensors   Science   Sharded   Synthetic data   Tensorflow
Model Card on HF 🤗: https://huggingface.co/Weyaxi/Einstein-v4-Qwen-1.5-32B 

Einstein V4 Qwen 1.5 32B Benchmarks

ARC: 62.37 vs 96.7 (so35)-35.5%
HellaSwag: 83.85 vs 95.3 (gpt4)-12%
MMLU: 74.04 vs 88.3 (so35)-16.1%
TruthfulQA: 58.86 vs 59 (gpt4)-0.2%
WinoGrande: 80.43 vs 87.5 (gpt4)-8.1%
GSM8K: 51.71 vs 96.4 (so35)-46.4%
LLME Score: 0.19485
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Einstein V4 Qwen 1.5 32B (Weyaxi/Einstein-v4-Qwen-1.5-32B)

Einstein V4 Qwen 1.5 32B Parameters and Internals

Model Type 
text generation, multimodal
Additional Notes 
The model is specifically tuned for science domains such as physics, chemistry, biology, and math using synthetic data and various datasets.
Training Details 
Data Sources:
allenai/ai2_arc, camel-ai/physics, camel-ai/chemistry, camel-ai/biology, camel-ai/math, metaeval/reclor, openbookqa, mandyyyyii/scibench, derek-thomas/ScienceQA, TIGER-Lab/ScienceEval, jondurbin/airoboros-3.2, LDJnr/Capybara, Cot-Alpaca-GPT4-From-OpenHermes-2.5, STEM-AI-mtl/Electrical-engineering, knowrohit07/saraswati-stem, sablo/oasst2_curated, glaiveai/glaive-code-assistant, lmsys/lmsys-chat-1m, TIGER-Lab/MathInstruct, bigbio/med_qa, meta-math/MetaMathQA-40K, openbookqa, piqa, metaeval/reclor, derek-thomas/ScienceQA, scibench, sciq, Open-Orca/SlimOrca, migtissera/Synthia-v1.3, TIGER-Lab/ScienceEval
Methodology:
Fine-tuning with QLoRA on diverse datasets
Context Length:
4096
Hardware Used:
8xRTX3090, 1xRTXA6000
Model Architecture:
QLoRA fine-tuned version of Qwen1.5
Input Output 
Input Format:
ChatML
Accepted Modalities:
text
Output Format:
text
Release Notes 
Version:
v4
Notes:
Full fine-tuned for 2 epochs. Total number of steps was 3352.
LLM NameEinstein V4 Qwen 1.5 32B
Repository 🤗https://huggingface.co/Weyaxi/Einstein-v4-Qwen-1.5-32B 
Base Model(s)  Qwen/Qwen1.5-32B   Qwen/Qwen1.5-32B
Model Size32b
Required VRAM64.6 GB
Updated2025-02-22
MaintainerWeyaxi
Model Typeqwen2
Model Files  4.9 GB: 1-of-14   4.8 GB: 2-of-14   4.8 GB: 3-of-14   4.8 GB: 4-of-14   4.8 GB: 5-of-14   4.8 GB: 6-of-14   4.8 GB: 7-of-14   4.8 GB: 8-of-14   4.8 GB: 9-of-14   4.8 GB: 10-of-14   4.8 GB: 11-of-14   4.8 GB: 12-of-14   4.8 GB: 13-of-14   2.1 GB: 14-of-14
Supported Languagesen
Model ArchitectureQwen2ForCausalLM
Licenseother
Context Length32768
Model Max Length32768
Transformers Version4.40.0.dev0
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Vocabulary Size152064
Torch Data Typebfloat16
Errorsreplace

Best Alternatives to Einstein V4 Qwen 1.5 32B

Best Alternatives
Context / RAM
Downloads
Likes
Openbuddy Qwq 32B V24.2 200K195K / 65.8 GB803
Openbuddy Qwq 32B V24.1 200K195K / 65.8 GB903
...y Qwen2.5coder 32B V24.1q 200K195K / 65.8 GB142
DeepSeek R1 Distill Qwen 32B128K / 65.7 GB9727681144
...k R1 Distill Qwen 32B Japanese128K / 65.8 GB16285239
Qwen2.5 32B128K / 65.5 GB72415104
Ultiima 32B128K / 65.8 GB4865
RomboUltima 32B128K / 20.7 GB691
...wen2.5 32B Inst BaseMerge TIES128K / 65.8 GB30710
...wen2.5 32B Inst BaseMerge TIES128K / 65.8 GB231
Note: green Score (e.g. "73.2") means that the model is better than Weyaxi/Einstein-v4-Qwen-1.5-32B.

Rank the Einstein V4 Qwen 1.5 32B 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? 43470 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.
Release v20241227
  Support LLM Explorer