WizardLM 2 7B
by Microsoft · wizardlm family
7B
parameters
text-generation code-generation reasoning creative-writing summarization
WizardLM 2 7B is Microsoft's instruction-tuned language model trained using the Evol-Instruct methodology, which progressively evolves training data to improve instruction following. It delivers strong performance across creative writing, code generation, and complex reasoning tasks for its size class. As the compact variant in the WizardLM 2 lineup, the 7B model offers an efficient option for local deployment on consumer hardware while maintaining competitive quality on benchmarks like MMLU and HumanEval.
Quick Start with Ollama
ollama run 7b-q4_K_M Resources
Ollama
| Creator | Microsoft |
| Parameters | 7B |
| Architecture | transformer-decoder |
| Context | 32K tokens |
| Released | Apr 15, 2024 |
| License | Apache 2.0 |
| Ollama | wizardlm2:7b |
Quantization Options
| Format | File Size | VRAM Required | Quality | Ollama Tag |
|---|---|---|---|---|
| Q4_K_M rec | 4.4 GB | 6.9 GB | | 7b-q4_K_M |
| Q8_0 | 7.4 GB | 9.9 GB | | 7b-q8_0 |
| F16 | 14 GB | 17 GB | | 7b-fp16 |
Compatible Hardware
Q4_K_M requires 6.9 GB VRAM
Benchmark Scores
65.0
mmlu