Arxiv:2010.04245 Arxiv:2108.12409 Arxiv:2205.14135 Composer Dataset:mosaicml/dolly hhrlhf Ggml Instruct Llm-foundry Mosaicml Mpt Quantized Region:us
MPT 7B Instruct GGML Benchmarks
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)
MPT 7B Instruct GGML Parameters and Internals
| LLM Name | MPT 7B Instruct GGML |
| Repository 🤗 | https://huggingface.co/TheBloke/MPT-7B-Instruct-GGML |
| Base Model(s) | |
| Model Size | 7b |
| Required VRAM | 3.7 GB |
| Updated | 2024-10-18 |
| Maintainer | TheBloke |
| Model Type | mpt |
| Instruction-Based | Yes |
| Model Files | |
| GGML Quantization | Yes |
| Quantization Type | ggml |
| Model Architecture | AutoModel |
| License | cc-by-sa-3.0 |
Best Alternatives to MPT 7B Instruct GGML
Best Alternatives |
Context / RAM |
Downloads |
Likes |
|---|---|---|---|
| CodeLlama 7B Instruct GGML | 0K / 3 GB | 13 | 20 |
| Llama 2 7B 32K Instruct GGML | 0K / 2.9 GB | 20 | 8 |
| Vigogne 2 7B Instruct GGML | 0K / 2.9 GB | 3 | 2 |
| Falcon 7B Instruct GGML | 0K / 4.1 GB | 93 | 42 |
| Redpajama 7B GGML | 0K / 3.9 GB | 69 | 3 |
| Mpt 7B GGML | 0K / 3.7 GB | 28 | 46 |
| Qwen 2.5 7B Rp Lora | 0K / 0.2 GB | 0 | 1 |
| Koishi 7B Qlora Gguf | 0K / 2.7 GB | 233 | 0 |
| Pixel | 8K / 4.4 GB | 63 | 0 |
| ... 7b 448 Qinstruct Preview V0.1 | 2K / 17.3 GB | 6 | 4 |
Note: green Score (e.g. "73.2") means that the model is better than TheBloke/MPT-7B-Instruct-GGML .
Rank the MPT 7B Instruct GGML 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? 36966 in total.
Original data from HuggingFace, OpenCompass and various public git repos.
Release v2024072803