Granite Code Models
Collection
A series of code models trained by IBM licensed under Apache 2.0 license. We release both the base pretrained and instruct models.
•
23 items
•
Updated
•
178
ibm-granite/granite-34b-code-base-8k-GGUF
This is the Q4_K_M converted version of the original ibm-granite/granite-34b-code-base-8k.
Refer to the original model card for more details.
Use with llama.cpp
git clone https://github.com/ggerganov/llama.cpp
cd llama.cpp
# install
make
# run generation
./main -m granite-34b-code-base-8k-GGUF/granite-34b-code-base.Q4_K_M.gguf -n 128 -p "def generate_random(x: int):" --color
- Downloads last month
- 112
Datasets used to train ibm-granite/granite-34b-code-base-8k-GGUF
Collection including ibm-granite/granite-34b-code-base-8k-GGUF
Evaluation results
- pass@1 on MBPPself-reported47.200
- pass@1 on MBPP+self-reported53.100
- pass@1 on HumanEvalSynthesis(Python)self-reported48.200
- pass@1 on HumanEvalSynthesis(JavaScript)self-reported54.900
- pass@1 on HumanEvalSynthesis(Java)self-reported61.600
- pass@1 on HumanEvalSynthesis(Go)self-reported40.200
- pass@1 on HumanEvalSynthesis(C++)self-reported50.000
- pass@1 on HumanEvalSynthesis(Rust)self-reported39.600
- pass@1 on HumanEvalExplain(Python)self-reported42.700
- pass@1 on HumanEvalExplain(JavaScript)self-reported26.200