metadata

license: apache-2.0
library_name: transformers
tags:
  - code
  - granite
  - llama-cpp
  - gguf-my-repo
base_model: ibm-granite/granite-34b-code-instruct
datasets:
  - bigcode/commitpackft
  - TIGER-Lab/MathInstruct
  - meta-math/MetaMathQA
  - glaiveai/glaive-code-assistant-v3
  - glaive-function-calling-v2
  - bugdaryan/sql-create-context-instruction
  - garage-bAInd/Open-Platypus
  - nvidia/HelpSteer
metrics:
  - code_eval
pipeline_tag: text-generation
inference: true
model-index:
  - name: granite-34b-code-instruct
    results:
      - task:
          type: text-generation
        dataset:
          name: HumanEvalSynthesis(Python)
          type: bigcode/humanevalpack
        metrics:
          - type: pass@1
            value: 62.2
            name: pass@1
          - type: pass@1
            value: 56.7
            name: pass@1
          - type: pass@1
            value: 62.8
            name: pass@1
          - type: pass@1
            value: 47.6
            name: pass@1
          - type: pass@1
            value: 57.9
            name: pass@1
          - type: pass@1
            value: 41.5
            name: pass@1
          - type: pass@1
            value: 53
            name: pass@1
          - type: pass@1
            value: 45.1
            name: pass@1
          - type: pass@1
            value: 50.6
            name: pass@1
          - type: pass@1
            value: 36
            name: pass@1
          - type: pass@1
            value: 42.7
            name: pass@1
          - type: pass@1
            value: 23.8
            name: pass@1
          - type: pass@1
            value: 54.9
            name: pass@1
          - type: pass@1
            value: 47.6
            name: pass@1
          - type: pass@1
            value: 55.5
            name: pass@1
          - type: pass@1
            value: 51.2
            name: pass@1
          - type: pass@1
            value: 47
            name: pass@1
          - type: pass@1
            value: 45.1
            name: pass@1

cobrakenji/granite-34b-code-instruct-Q4_K_M-GGUF

This model was converted to GGUF format from ibm-granite/granite-34b-code-instruct 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 --hf-repo cobrakenji/granite-34b-code-instruct-Q4_K_M-GGUF --hf-file granite-34b-code-instruct-q4_k_m.gguf -p "The meaning to life and the universe is"

Server:

llama-server --hf-repo cobrakenji/granite-34b-code-instruct-Q4_K_M-GGUF --hf-file granite-34b-code-instruct-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.

./main --hf-repo cobrakenji/granite-34b-code-instruct-Q4_K_M-GGUF --hf-file granite-34b-code-instruct-q4_k_m.gguf -p "The meaning to life and the universe is"

./server --hf-repo cobrakenji/granite-34b-code-instruct-Q4_K_M-GGUF --hf-file granite-34b-code-instruct-q4_k_m.gguf -c 2048