← Run Locally · Wonke amamodeli
ngu-meta-llama
Llama 3.3 70B Instruct
meta-llama/Llama-3.3-70B-Instruct
Wonke amanani ayizilinganiso. Ukusetshenziswa kwangempela kwenkumbulo nejubane kuncike kumadrayivha enu, ezilungiselelweni nakwezinye izinhlelo ezisebenzayo.
- Amapharamitha
- 71B
- Ilayisense
- llama3.3
- Ishicilelwe
- Nov 26, 2024
Ilayisense ngeyomkhiqizi wemodeli. Hlolani imigomo yayo ngaphambi kokusebenzisa imodeli.
I-Hugging Face iyanicela ukuthi nemukele imigomo yomkhiqizi ngaphambi kokulanda le modeli. Akukho khithi elithengiswayo yayo.
Okwasekuqaleni ku-Hugging Face ↗Inkumbulo edingekayo nge-quantization
Ngomongo wama-token angu-8,192, kuhlanganisa ne-0.8 GB ye-runtime.
| I-quantization | Ifayela lemodeli | Isamba |
|---|---|---|
| Q4_K_M | 42.8 GB | 47.9 GB(okulinganiselwe) |
| Q5_K_M | 50.2 GB | 56.0 GB(okulinganiselwe) |
| Q6_K | 57.9 GB | 64.4 GB(okulinganiselwe) |
| Q8_0 | 75.0 GB | 83.3 GB(okulinganiselwe) |
| F16 | 141.1 GB | 156.0 GB(okulinganiselwe) |
I-hardware eyisebenzisa ekhadini lezithombe
Ku-Q4_K_M nomongo wama-token angu-8,192.
Akukho kuma-hardware esiwafake ohlwini okuyisebenzisa ngokuphelele ekhadini lezithombe kule zilungiselelo.
Hlolani ekhompyutheni yenuImodeli efanayo nge-API
Umthengisi ophansi kakhulu: NanoGPT, $0.99 ngenyanga ngokusetshenziswa okuvamile (izicelo ezingu-200 ngosuku).
Bhekani amanani ku-Model Price MapImininingwane yamamodeli isuka ohlwini lomphakathi lwe-Hugging Face (amagama, osayizi, amalayisense). Ikhasi ngalinye lemodeli lixhumanisa nelokuqala.