GGML files are for CPU + GPU inference using llama.cpp
How to run in llama.cpp
./main -t 10 -ngl 32 -m ggml-model-q8_0.bin --color -c 2048 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "### Instruction: Write JQL(Jira query Language) for give input ### Input: stories assigned to manthan which are created in last 10 days with highest priority and label is set to release ### Response:"
Change -t 10 to the number of physical CPU cores you have. For example if your system has 8 cores/16 threads, use -t 8.
Change -ngl 32 to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
Tto have a chat-style conversation, replace the -p <PROMPT> argument with -i -ins
How to run in text-generation-webui
Further instructions here: text-generation-webui/docs/llama.cpp-models.md.
How to run using LangChain
Instalation on CPU
pip install llama-cpp-python
Instalation on GPU
CMAKE_ARGS="-DLLAMA_CUBLAS=on" FORCE_CMAKE=1 pip install llama-cpp-python
from langchain.llms import LlamaCpp
from langchain import PromptTemplate, LLMChain
from langchain.callbacks.manager import CallbackManager
from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
n_gpu_layers = 40 # Change this value based on your model and your GPU VRAM pool.
n_batch = 512 # Should be between 1 and n_ctx, consider the amount of VRAM in your GPU.
n_ctx=2048
callback_manager = CallbackManager([StreamingStdOutCallbackHandler()])
# Make sure the model path is correct for your system!
llm = LlamaCpp(
model_path="./ggml-model-q8_0.bin",
n_gpu_layers=n_gpu_layers, n_batch=n_batch,
callback_manager=callback_manager,
verbose=True,
n_ctx=n_ctx
)
llm("""### Instruction:
Write JQL(Jira query Language) for give input
### Input:
stories assigned to manthan which are created in last 10 days with highest priority and label is set to release
### Response:""")
For more information refer LangChain