Adapter Alignment-handbook Base model:adapter:microsoft/p... Base model:microsoft/phi-2 Dataset:huggingfaceh4/ultrafee... Dpo Finetuned Generated from trainer Lora Peft Phi Region:us Safetensors Tensorboard Trl
Model Card on HF 🤗: https://huggingface.co/DUAL-GPO/phi-2-gpo-renew2-b0.001-0.5ultrafeedback-i1
Phi 2 Gpo Renew2 B0.001 0.5ultrafeedback I1 Benchmarks
ARC: 60.92 vs 96.7 (so35)-37%
HellaSwag: 75.02 vs 95.3 (gpt4)-21.3%
MMLU: 57.97 vs 88.3 (so35)-34.3%
TruthfulQA: 44.36 vs 59 (gpt4)-24.8%
WinoGrande: 73.95 vs 87.5 (gpt4)-15.5%
GSM8K: 55.04 vs 96.4 (so35)-42.9%
LLME Score: 0.20172
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
What is the LLM Explorer Rank (Score)
Phi 2 Gpo Renew2 B0.001 0.5ultrafeedback I1 Parameters and Internals
| Training Details |
|
| LLM Name | Phi 2 Gpo Renew2 B0.001 0.5ultrafeedback I1 |
| Repository 🤗 | https://huggingface.co/DUAL-GPO/phi-2-gpo-renew2-b0.001-0.5ultrafeedback-i1 |
| Base Model(s) | |
| Required VRAM | 0 GB |
| Updated | 2024-11-17 |
| Maintainer | DUAL-GPO |
| Model Files | |
| Model Architecture | Adapter |
| License | mit |
| Model Max Length | 2048 |
| Is Biased | none |
| Tokenizer Class | CodeGenTokenizer |
| Padding Token | <|endoftext|> |
| PEFT Type | LORA |
| LoRA Model | Yes |
| PEFT Target Modules | v_proj|q_proj|dense|k_proj |
| LoRA Alpha | 128 |
| LoRA Dropout | 0.05 |
| R Param | 128 |
Best Alternatives to Phi 2 Gpo Renew2 B0.001 0.5ultrafeedback I1
Best Alternatives |
Context / RAM |
Downloads |
Likes |
|---|---|---|---|
| Phi 3 Mini 4K Instruct Sa V0.1 | 0K / 0 GB | 8 | 0 |
| Samantha Omni Humanlike Lora | 0K / 0 GB | 56 | 3 |
| ...is Violet Toxic GRPO V0.4 Lora | 0K / 0.5 GB | 12 | 0 |
| Reflection Model | 0K / 0.2 GB | 0 | 1 |
| SpectraMind | 0K / 16.1 GB | 120 | 3 |
| ...mall Physics Finetuned Adapter | 0K / 0.1 GB | 8 | 1 |
| SpectraMindQ | 0K / 0.2 GB | 8 | 1 |
| L3.1 Spark R64 LoRA | 0K / 0.4 GB | 8 | 0 |
| Mistral Small Fujin Qlora | 0K / 0.8 GB | 47 | 2 |
| Mistral Small Dampf Qlora | 0K / 0.8 GB | 19 | 0 |
Note: green Score (e.g. "73.2") means that the model is better than DUAL-GPO/phi-2-gpo-renew2-b0.001-0.5ultrafeedback-i1 .
Rank the Phi 2 Gpo Renew2 B0.001 0.5ultrafeedback I1 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! 🌟
What open-source LLMs or SLMs are you in search of? 43470 in total.
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241227