README.md

metadata

license: apache-2.0
datasets:
  - bigcode/the-stack
  - bigcode/the-stack-v2
  - bigcode/starcoderdata
  - bigcode/commitpack
library_name: transformers
tags:
  - code
  - llama-cpp
  - gguf-my-repo
base_model: JetBrains/Mellum-4b-sft-all
model-index:
  - name: Mellum-4b-sft-all
    results:
      - task:
          type: text-generation
        dataset:
          name: RepoBench 1.1 (Python)
          type: tianyang/repobench_python_v1.1
        metrics:
          - type: exact_match
            value: 0.2823
            name: EM
            verified: false
          - type: exact_match
            value: 0.287
            name: EM ≤ 8k
            verified: false
          - type: exact_match
            value: 0.2638
            name: EM
            verified: false
          - type: exact_match
            value: 0.293
            name: EM
            verified: false
          - type: exact_match
            value: 0.3042
            name: EM
            verified: false
          - type: exact_match
            value: 0.2685
            name: EM
            verified: false
          - type: exact_match
            value: 0.2818
            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: 0.2867
            name: EM
            verified: false
          - type: exact_match
            value: 0.3023
            name: EM ≤ 8k
            verified: false
          - type: exact_match
            value: 0.2883
            name: EM
            verified: false
          - type: exact_match
            value: 0.3228
            name: EM
            verified: false
          - type: exact_match
            value: 0.2958
            name: EM
            verified: false
          - type: exact_match
            value: 0.2447
            name: EM
            verified: false
          - type: exact_match
            value: 0.2821
            name: EM
            verified: false
      - task:
          type: text-generation
        dataset:
          name: SAFIM
          type: gonglinyuan/safim
        metrics:
          - type: pass@1
            value: 0.5285
            name: pass@1
            verified: false
          - type: pass@1
            value: 0.6548
            name: pass@1
            verified: false
          - type: pass@1
            value: 0.4005
            name: pass@1
            verified: false
          - type: pass@1
            value: 0.5303
            name: pass@1
            verified: false
      - task:
          type: text-generation
        dataset:
          name: HumanEval Infilling (Single-Line)
          type: loubnabnl/humaneval_infilling
        metrics:
          - type: pass@1
            value: 0.8083
            name: pass@1
            verified: false
          - type: pass@1
            value: 0.4819
            name: pass@1
            verified: false
          - type: pass@1
            value: 0.372
            name: pass@1
            verified: false
          - type: pass@1
            value: 0.4024
            name: pass@1
            verified: false

LLukas22/Mellum-4b-sft-all-Q4_K_M-GGUF

This model was converted to GGUF format from JetBrains/Mellum-4b-sft-all using llama.cpp via the ggml.ai's GGUF-my-repo space. Refer to the original model card for more details on the model.

Use with llama.cpp

Install llama.cpp through brew (works on Mac and Linux)

brew install llama.cpp

Invoke the llama.cpp server or the CLI.

CLI:

llama-cli --hf-repo LLukas22/Mellum-4b-sft-all-Q4_K_M-GGUF --hf-file mellum-4b-sft-all-q4_k_m.gguf -p "The meaning to life and the universe is"

Server:

llama-server --hf-repo LLukas22/Mellum-4b-sft-all-Q4_K_M-GGUF --hf-file mellum-4b-sft-all-q4_k_m.gguf -c 2048

Note: You can also use this checkpoint directly through the usage steps listed in the Llama.cpp repo as well.

Step 1: Clone llama.cpp from GitHub.

git clone https://github.com/ggerganov/llama.cpp

Step 2: Move into the llama.cpp folder and build it with LLAMA_CURL=1 flag along with other hardware-specific flags (for ex: LLAMA_CUDA=1 for Nvidia GPUs on Linux).

cd llama.cpp && LLAMA_CURL=1 make

Step 3: Run inference through the main binary.

./llama-cli --hf-repo LLukas22/Mellum-4b-sft-all-Q4_K_M-GGUF --hf-file mellum-4b-sft-all-q4_k_m.gguf -p "The meaning to life and the universe is"

or

./llama-server --hf-repo LLukas22/Mellum-4b-sft-all-Q4_K_M-GGUF --hf-file mellum-4b-sft-all-q4_k_m.gguf -c 2048