Adapter Alignment-handbook Base model:adapter:meta-llama/... Base model:meta-llama/meta-lla... Dataset:huggingfaceh4/ultrafee... Finetuned Generated from trainer Instruct Llama Lora Orpo Peft Region:us Safetensors Trl
Model Card on HF 🤗: https://huggingface.co/statking/Meta-Llama-3-8B-Instruct-ORPO-QLoRA
Meta Llama 3 8B Instruct ORPO QLoRA Benchmarks
ARC: 58.19 vs 96.7 (so35)-39.8%
HellaSwag: 79.42 vs 95.3 (gpt4)-16.7%
MMLU: 65.59 vs 88.3 (so35)-25.7%
TruthfulQA: 48.38 vs 59 (gpt4)-18%
WinoGrande: 76.56 vs 87.5 (gpt4)-12.5%
GSM8K: 58.61 vs 96.4 (so35)-39.2%
LLME Score: 0.21379
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)
Meta Llama 3 8B Instruct ORPO QLoRA Parameters and Internals
| Training Details |
|
| LLM Name | Meta Llama 3 8B Instruct ORPO QLoRA |
| Repository 🤗 | https://huggingface.co/statking/Meta-Llama-3-8B-Instruct-ORPO-QLoRA |
| Base Model(s) | |
| Model Size | 8b |
| Required VRAM | 0.7 GB |
| Updated | 2024-11-17 |
| Maintainer | statking |
| Instruction-Based | Yes |
| Model Files | |
| Model Architecture | Adapter |
| License | llama3 |
| Model Max Length | 2048 |
| Is Biased | none |
| Tokenizer Class | PreTrainedTokenizerFast |
| Padding Token | <|eot_id|> |
| PEFT Type | LORA |
| LoRA Model | Yes |
| PEFT Target Modules | up_proj|v_proj|k_proj|o_proj|gate_proj|q_proj|down_proj |
| LoRA Alpha | 128 |
| LoRA Dropout | 0.05 |
| R Param | 128 |
Best Alternatives to Meta Llama 3 8B Instruct ORPO QLoRA
Best Alternatives |
Context / RAM |
Downloads |
Likes |
|---|---|---|---|
| ... 3 8B Instruct Bvr Finetune V3 | 8K / 16.1 GB | 5 | 0 |
| Flippa V6 | 0K / 0 GB | 11 | 1 |
| ...lama 3 1 8B Instruct Orca ORPO | 0K / 0.1 GB | 126 | 2 |
| ...ight AGI Llama3.1 8B UC200k V2 | 0K / 0.2 GB | 88 | 2 |
| Vortex2 | 0K / 4.4 GB | 8 | 0 |
| FineLlama3.1 8B Instruct Lora | 0K / 0.2 GB | 0 | 1 |
| Llama3 8B Instruct Code | 0K / 0.2 GB | 2 | 1 |
| Llama 3 8B Claudstruct V3 | 0K / 0.1 GB | 11 | 0 |
| Llama 3 8B Claudstruct V1 | 0K / 0.1 GB | 7 | 0 |
| Llama 3 8B Instruct 80K QLoRA | 0K / 2.2 GB | 0 | 24 |
Note: green Score (e.g. "73.2") means that the model is better than statking/Meta-Llama-3-8B-Instruct-ORPO-QLoRA .
Rank the Meta Llama 3 8B Instruct ORPO QLoRA 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? 40066 in total.
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241217