[whatmodel.sh](../../README.md) / Pythia 2.8B
# Pythia 2.8B
EleutherAI · released 2023-02 · 2.78B params · apache-2.0
**score 20.1** ±1.4 · #701 of 732 · from 15 benchmark results across 5 sources
Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only.
## What it's good at
● marks Pythia 2.8B among every model on that task or test, worst to best.
other tests (12)
test result vs the others in its run difficulty counts why
WinoGrande (maker) 59.4% ▄····▄▆▂▄●▆█ beats 11/19 easy · avg 87% 19% several runs
ARC Challenge (maker) 29.5% █··▁·▁▁·●▂▁▂ beats 13/19 medium · avg 71% 14% sharp · too hard for it
ARC Easy (maker) 64.4% ▃·▅▃▃▅·██●██ beats 13/19 easy · avg 87% 13% blunt · at its level
ImageNetVC visual commonsense, te… 44.1% ▂·▄·▂█●▄█▄▂▆ beats 8/21 easy · avg 74% 12% blunt · at its level
PIQA (maker) 73.9% ▂·▄▄·▄▂▆▂●█▄ beats 13/19 easy · avg 87% 10% blunt · at its level · several runs
LAMBADA (OpenAI version), last-wo… 64.7% ▃··▅▃▃▅▅▅●██ beats 13/19 very easy · avg 87% 8.9% blunt · at its level · several runs
SciQ (maker) 88.2% ▂··▂▃▂▂▆▃●█▂ beats 12/19 very easy · avg 98% 7.5% blunt
HellaSwag 59.3% ▁▁▁▂▂▂●▆█▇▄▂ beats 38/135 very easy · avg 89% 5.4% at its level · small run
MedQA (USMLE, 4 options) 24.3% ●▁··▄▁▁··▄▃█ beats 0/17 medium · avg 71% 3.6% sharp · too hard for it
MedMCQA 27.2% ▂●··▄▄▂▂·▄██ beats 1/17 medium · avg 63% 2.7% typical
Open PT-LLM Leaderboard average (… 24.0% ●▁▁▁▁▁▁▁▂█▅▁ beats 8/276 medium · avg 72% 1.6% noisy
PubMedQA 64.0% ▂·▂▆·●▂▄▄▂█▂ beats 5/17 medium · avg 75% 0.9% blunt · at its level
source results share of the score
Pythia: A Suite for Analyzing Large Lan… 6 62% ▇▇▇▇▇▇▇
RedPajama 4 17% ▇▇
ImageNetVC: Zero- and Few-Shot Visual C… 1 12% ▇
Open Medical-LLM Leaderboard results 3 7% ▇
Open Portuguese LLM Leaderboard 1 2% ▇
prior: pull toward the middle 1% ▇
weighted average = 20.1
Licences: Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling ([CC BY-SA 4.0](../../licenses/paper-2304.01373v2-CC-BY-SA-4.0.txt)) · RedPajama ([CC BY 4.0](../../licenses/redpajama-2024-original-zero-shot-CC-BY-4.0.txt)) · ImageNetVC: Zero- and Few-Shot Visual Commonsense Evaluation on 1000 ImageNet Categories ([CC BY 4.0](../../licenses/acl-2023-findings-emnlp-133-CC-BY-4-0.txt)) · Open Medical-LLM Leaderboard results ([Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0)) · Open Portuguese LLM Leaderboard ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt))
## How to run it
### llama.cpp (recommended)
1. Install llama.cpp. Homebrew and winget installs name the commands llama-cli and llama-server instead. [Other ways to install](https://github.com/ggml-org/llama.cpp/blob/master/docs/install.md)
```sh
# macOS / Linux
curl -LsSf https://llama.app/install.sh | sh
# Windows
irm https://llama.app/install.ps1 | iex
```
2. Run Pythia 2.8B (Q4_K_M, needs ~1.8 GB, from [huggingface.co/mradermacher](https://huggingface.co/mradermacher/pythia-2.8b-GGUF/tree/147586a405b950ff991ba909a3193ffafa98b014))
```sh
llama cli -hf mradermacher/pythia-2.8b-GGUF:Q4_K_M
```
3. Or start a local server and open the address it prints (http://localhost:9931)
```sh
llama serve -hf mradermacher/pythia-2.8b-GGUF:Q4_K_M
```
Other versions: Q8_0 (~3 GB) `llama cli -hf mradermacher/pythia-2.8b-GGUF:Q8_0`
### LM Studio
1. Install LM Studio. Open the app once before using the lms command. [Download it (macOS / Windows / Linux)](https://lmstudio.ai/download)
2. Run Pythia 2.8B (Q4_K_M, needs ~1.8 GB, from [huggingface.co/mradermacher](https://huggingface.co/mradermacher/pythia-2.8b-GGUF/tree/147586a405b950ff991ba909a3193ffafa98b014))
```sh
lms get https://huggingface.co/mradermacher/pythia-2.8b-GGUF@Q4_K_M -y
```
3. Then run lms ls to see its name, and lms chat followed by that name
Other versions: Q8_0 (~3 GB) `lms get https://huggingface.co/mradermacher/pythia-2.8b-GGUF@Q8_0 -y`
### Ollama
1. Install Ollama. [Download it (macOS / Windows)](https://ollama.com/download)
```sh
# Linux
curl -fsSL https://ollama.com/install.sh | sh
```
2. Run Pythia 2.8B (Q4_K_M, needs ~1.8 GB, from [huggingface.co/mradermacher](https://huggingface.co/mradermacher/pythia-2.8b-GGUF/tree/147586a405b950ff991ba909a3193ffafa98b014))
```sh
ollama run hf.co/mradermacher/pythia-2.8b-GGUF:Q4_K_M
```
Other versions: Q8_0 (~3 GB) `ollama run hf.co/mradermacher/pythia-2.8b-GGUF:Q8_0`
### Other ways to run it
1. Run it as published (fp16, needs ~5.68 GB, from [huggingface.co/EleutherAI](https://huggingface.co/EleutherAI/pythia-2.8b/tree/2a259cdd96a4beb1cdf467512e3904197345f6a9))
```sh
uv run --with torch --with 'transformers<5' python -c "from transformers import pipeline; print(pipeline('text-generation', model='EleutherAI/pythia-2.8b')('Once upon a time', max_new_tokens=40)[0]['generated_text'])"
```
Model repository and documentation: [huggingface.co/EleutherAI](https://huggingface.co/EleutherAI/pythia-2.8b)
[interactive version](https://whatmodel.sh/model/pythia-2-8b) · [JSON](https://whatmodel.sh/api?format=json&all=1)