Hugging Face's logo

tensorblock
/
granite-3b-code-base-128k-GGUF

Text Generation
Transformers
GGUF
code
granite
TensorBlock
GGUF
Eval Results
Model card Files Community
granite-3b-code-base-128k-GGUF / README.md
morriszms's picture
morriszms
Update README.md
63c8a7d verified
preview code
|
raw
history blame contribute delete
8.31 kB
metadata
pipeline_tag: text-generation
inference: false
license: apache-2.0
datasets:
  - codeparrot/github-code-clean
  - bigcode/starcoderdata
  - open-web-math/open-web-math
  - math-ai/StackMathQA
metrics:
  - code_eval
library_name: transformers
tags:
  - code
  - granite
  - TensorBlock
  - GGUF
base_model: ibm-granite/granite-3b-code-base-128k
model-index:
  - name: granite-3b-code-base-128k
    results:
      - task:
          type: text-generation
        dataset:
          name: HumanEvalSynthesis (Python)
          type: bigcode/humanevalpack
        metrics:
          - type: pass@1
            value: 36
            name: pass@1
            verified: false
          - type: pass@1
            value: 30.5
            name: pass@1
            verified: false
          - type: pass@1
            value: 22.4
            name: pass@1
            verified: false
          - type: pass@1
            value: 19.9
            name: pass@1
            verified: false
      - task:
          type: text-generation
        dataset:
          name: RepoQA (Python@16K)
          type: repoqa
        metrics:
          - type: pass@1 (thresh=0.5)
            value: 40
            name: pass@1 (thresh=0.5)
            verified: false
          - type: pass@1 (thresh=0.5)
            value: 36
            name: pass@1 (thresh=0.5)
            verified: false
          - type: pass@1 (thresh=0.5)
            value: 37
            name: pass@1 (thresh=0.5)
            verified: false
          - type: pass@1 (thresh=0.5)
            value: 27
            name: pass@1 (thresh=0.5)
            verified: false
          - type: pass@1 (thresh=0.5)
            value: 29
            name: pass@1 (thresh=0.5)
            verified: false
      - task:
          type: text-generation
        dataset:
          name: LCC (Balanced)
          type: lcc
        metrics:
          - type: Exact Match@4K
            value: 54.6
            name: Exact Match@4K
            verified: false
          - type: Exact Match@8K
            value: 56.8
            name: Exact Match@8K
            verified: false
          - type: Exact Match@16K
            value: 52.2
            name: Exact Match@16K
            verified: false
          - type: Exact Match@32K
            value: 57.8
            name: Exact Match@32K
            verified: false
      - task:
          type: text-generation
        dataset:
          name: RepoBench-P (Balanced)
          type: repobench
        metrics:
          - type: Exact Match@4K
            value: 39.8
            name: Exact Match@4K
            verified: false
          - type: Exact Match@8K
            value: 46.8
            name: Exact Match@8K
            verified: false
          - type: Exact Match@16K
            value: 43.1
            name: Exact Match@16K
            verified: false
          - type: Exact Match@32K
            value: 45.3
            name: Exact Match@32K
            verified: false
TensorBlock

Feedback and support: TensorBlock's Twitter/X, Telegram Group and Discord server

ibm-granite/granite-3b-code-base-128k - GGUF

This repo contains GGUF format model files for ibm-granite/granite-3b-code-base-128k.

The files were quantized using machines provided by TensorBlock, and they are compatible with llama.cpp as of commit b4011.

Our projects

Awesome MCP Servers TensorBlock Studio
Project A Project B
A comprehensive collection of Model Context Protocol (MCP) servers. A lightweight, open, and extensible multi-LLM interaction studio.
👀 See what we built 👀 👀 See what we built 👀
## Prompt template

Model file specification

Filename Quant type File Size Description
granite-3b-code-base-128k-Q2_K.gguf Q2_K 1.339 GB smallest, significant quality loss - not recommended for most purposes
granite-3b-code-base-128k-Q3_K_S.gguf Q3_K_S 1.552 GB very small, high quality loss
granite-3b-code-base-128k-Q3_K_M.gguf Q3_K_M 1.727 GB very small, high quality loss
granite-3b-code-base-128k-Q3_K_L.gguf Q3_K_L 1.876 GB small, substantial quality loss
granite-3b-code-base-128k-Q4_0.gguf Q4_0 1.997 GB legacy; small, very high quality loss - prefer using Q3_K_M
granite-3b-code-base-128k-Q4_K_S.gguf Q4_K_S 2.014 GB small, greater quality loss
granite-3b-code-base-128k-Q4_K_M.gguf Q4_K_M 2.132 GB medium, balanced quality - recommended
granite-3b-code-base-128k-Q5_0.gguf Q5_0 2.417 GB legacy; medium, balanced quality - prefer using Q4_K_M
granite-3b-code-base-128k-Q5_K_S.gguf Q5_K_S 2.417 GB large, low quality loss - recommended
granite-3b-code-base-128k-Q5_K_M.gguf Q5_K_M 2.486 GB large, very low quality loss - recommended
granite-3b-code-base-128k-Q6_K.gguf Q6_K 2.862 GB very large, extremely low quality loss
granite-3b-code-base-128k-Q8_0.gguf Q8_0 3.706 GB very large, extremely low quality loss - not recommended

Downloading instruction

Command line

Firstly, install Huggingface Client

pip install -U "huggingface_hub[cli]"

Then, downoad the individual model file the a local directory

huggingface-cli download tensorblock/granite-3b-code-base-128k-GGUF --include "granite-3b-code-base-128k-Q2_K.gguf" --local-dir MY_LOCAL_DIR

If you wanna download multiple model files with a pattern (e.g., *Q4_K*gguf), you can try:

huggingface-cli download tensorblock/granite-3b-code-base-128k-GGUF --local-dir MY_LOCAL_DIR --local-dir-use-symlinks False --include='*Q4_K*gguf'