Arxiv:1910.09700 4-bit Autotrain compatible Endpoints compatible Gpt neox Gptq Quantized Region:us Safetensors Sharded Tensorflow
Model Card on HF 🤗: https://huggingface.co/Xu-Ouyang/pythia-12b-deduped-int4-step36000-GPTQ-wikitext2
Pythia 12B Deduped Int4 Step36000 GPTQ Wikitext2 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)
Pythia 12B Deduped Int4 Step36000 GPTQ Wikitext2 Parameters and Internals
| LLM Name | Pythia 12B Deduped Int4 Step36000 GPTQ Wikitext2 |
| Repository 🤗 | https://huggingface.co/Xu-Ouyang/pythia-12b-deduped-int4-step36000-GPTQ-wikitext2 |
| Model Size | 12b |
| Required VRAM | 6.9 GB |
| Updated | 2025-02-22 |
| Maintainer | Xu-Ouyang |
| Model Type | gpt_neox |
| Model Files | |
| GPTQ Quantization | Yes |
| Quantization Type | gptq |
| Model Architecture | GPTNeoXForCausalLM |
| Context Length | 2048 |
| Model Max Length | 2048 |
| Transformers Version | 4.41.2 |
| Tokenizer Class | GPTNeoXTokenizer |
| Vocabulary Size | 50688 |
| Torch Data Type | float16 |
Best Alternatives to Pythia 12B Deduped Int4 Step36000 GPTQ Wikitext2
Best Alternatives |
Context / RAM |
Downloads |
Likes |
|---|---|---|---|
| ... Int3 Step71000 GPTQ Wikitext2 | 2K / 5.5 GB | 76 | 0 |
| ...Int4 Step107000 GPTQ Wikitext2 | 2K / 6.9 GB | 76 | 0 |
| Merlyn Education Safety GPTQ | 2K / 6.9 GB | 61 | 3 |
| Dolly V2 12B Sharded 8bit | 2K / 12.4 GB | 19 | 4 |
| ...sst Sft 4 Pythia 12B Epoch 3.5 | 2K / 23.8 GB | 426081 | 364 |
| Dolly V2 12B | 2K / 23.8 GB | 6765 | 1953 |
| Pythia 12B | 2K / 23.8 GB | 14160 | 135 |
| Oasst Sft 1 Pythia 12B | 2K / 23.8 GB | 5010 | 278 |
| Pythia 12B Deduped | 2K / 23.8 GB | 9925 | 51 |
| Pythia 12B Sft V8 7K Steps | 2K / 23.8 GB | 2142 | 21 |
Note: green Score (e.g. "73.2") means that the model is better than Xu-Ouyang/pythia-12b-deduped-int4-step36000-GPTQ-wikitext2 .
Rank the Pythia 12B Deduped Int4 Step36000 GPTQ Wikitext2 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