README.md

metadata

license: apache-2.0
datasets:
  - bigcode/the-stack
  - bigcode/the-stack-v2
  - bigcode/starcoderdata
  - bigcode/commitpack
library_name: mlx
tags:
  - code
  - mlx
base_model: JetBrains/Mellum-4b-base
pipeline_tag: text-generation
model-index:
  - name: Mellum-4b-base
    results:
      - task:
          type: text-generation
        dataset:
          name: RepoBench 1.1 (Python)
          type: tianyang/repobench_python_v1.1
        metrics:
          - type: exact_match
            value: 0.2591
            name: EM
            verified: false
          - type: exact_match
            value: 0.2797
            name: EM ≤ 8k
            verified: false
          - type: exact_match
            value: 0.282
            name: EM
            verified: false
          - type: exact_match
            value: 0.2795
            name: EM
            verified: false
          - type: exact_match
            value: 0.2777
            name: EM
            verified: false
          - type: exact_match
            value: 0.2453
            name: EM
            verified: false
          - type: exact_match
            value: 0.211
            name: EM
            verified: false
      - task:
          type: text-generation
        dataset:
          name: RepoBench 1.1 (Java)
          type: tianyang/repobench_java_v1.1
        metrics:
          - type: exact_match
            value: 0.2858
            name: EM
            verified: false
          - type: exact_match
            value: 0.3108
            name: EM ≤ 8k
            verified: false
          - type: exact_match
            value: 0.3202
            name: EM
            verified: false
          - type: exact_match
            value: 0.3212
            name: EM
            verified: false
          - type: exact_match
            value: 0.291
            name: EM
            verified: false
          - type: exact_match
            value: 0.2492
            name: EM
            verified: false
          - type: exact_match
            value: 0.2474
            name: EM
            verified: false
      - task:
          type: text-generation
        dataset:
          name: SAFIM
          type: gonglinyuan/safim
        metrics:
          - type: pass@1
            value: 0.3811
            name: pass@1
            verified: false
          - type: pass@1
            value: 0.253
            name: pass@1
            verified: false
          - type: pass@1
            value: 0.3839
            name: pass@1
            verified: false
          - type: pass@1
            value: 0.5065
            name: pass@1
            verified: false
      - task:
          type: text-generation
        dataset:
          name: HumanEval Infilling (Single-Line)
          type: loubnabnl/humaneval_infilling
        metrics:
          - type: pass@1
            value: 0.6621
            name: pass@1
            verified: false
          - type: pass@1
            value: 0.3852
            name: pass@1
            verified: false
          - type: pass@1
            value: 0.2969
            name: pass@1
            verified: false

mlx-community/Mellum-4b-base-4bit

This model mlx-community/Mellum-4b-base-4bit was converted to MLX format from JetBrains/Mellum-4b-base using mlx-lm version 0.25.2.

Use with mlx

pip install mlx-lm

from mlx_lm import load, generate

model, tokenizer = load("mlx-community/Mellum-4b-base-4bit")

prompt = "hello"

if tokenizer.chat_template is not None:
    messages = [{"role": "user", "content": prompt}]
    prompt = tokenizer.apply_chat_template(
        messages, add_generation_prompt=True
    )

response = generate(model, tokenizer, prompt=prompt, verbose=True)