🎩 Magicoder is a family of 7B parameter models trained on 75K synthetic instruction data using OSS-Instruct, a novel approach to enlightening LLMs with open-source code snippets.
ollama run magicoder
Ollama kurulu değil mi? ollama.com/download — Windows, macOS ve Linux için ücretsiz. İlk çalıştırmada model indirilir, sonrası tamamen çevrimdışıdır.
Boyut büyüdükçe kalite artar, donanım ihtiyacı yükselir. Başlangıç için küçük varyantı deneyin.
| Etiket | Boyut | Bağlam | Girdi | Komut |
|---|---|---|---|---|
| latest | 3.8GB | 16K | Text |
ollama run magicoder:latest
|
| 7b | 3.8GB | 16K | Text |
ollama run magicoder:7b
|
| 7b-s-cl | 3.8GB | 16K | Text |
ollama run magicoder:7b-s-cl
|
| 7b-s-cl-q2_K | 2.8GB | 16K | Text |
ollama run magicoder:7b-s-cl-q2_K
|
| 7b-s-cl-q3_K_S | 2.9GB | 16K | Text |
ollama run magicoder:7b-s-cl-q3_K_S
|
| 7b-s-cl-q3_K_M | 3.3GB | 16K | Text |
ollama run magicoder:7b-s-cl-q3_K_M
|
| 7b-s-cl-q3_K_L | 3.6GB | 16K | Text |
ollama run magicoder:7b-s-cl-q3_K_L
|
| 7b-s-cl-q4_0 | 3.8GB | 16K | Text |
ollama run magicoder:7b-s-cl-q4_0
|
| 7b-s-cl-q4_1 | 4.2GB | 16K | Text |
ollama run magicoder:7b-s-cl-q4_1
|
| 7b-s-cl-q4_K_S | 3.9GB | 16K | Text |
ollama run magicoder:7b-s-cl-q4_K_S
|
| 7b-s-cl-q4_K_M | 4.1GB | 16K | Text |
ollama run magicoder:7b-s-cl-q4_K_M
|
| 7b-s-cl-q5_0 | 4.7GB | 16K | Text |
ollama run magicoder:7b-s-cl-q5_0
|
| 7b-s-cl-q5_1 | 5.1GB | 16K | Text |
ollama run magicoder:7b-s-cl-q5_1
|
| 7b-s-cl-q5_K_S | 4.7GB | 16K | Text |
ollama run magicoder:7b-s-cl-q5_K_S
|
| 7b-s-cl-q5_K_M | 4.8GB | 16K | Text |
ollama run magicoder:7b-s-cl-q5_K_M
|
| 7b-s-cl-q6_K | 5.5GB | 16K | Text |
ollama run magicoder:7b-s-cl-q6_K
|
| 7b-s-cl-q8_0 | 7.2GB | 16K | Text |
ollama run magicoder:7b-s-cl-q8_0
|
| 7b-s-cl-fp16 | 13GB | 16K | Text |
ollama run magicoder:7b-s-cl-fp16
|
Magicoder is a model family empowered by OSS-Instruct, a novel approach to enlightening LLMs with open-source code snippets for generating low-bias and high-quality instruction data for code.
OSS-Instruct mitigates the inherent bias of the LLM-synthesized instruction data by empowering them with a wealth of open-source references to produce more diverse, realistic, and controllable data.