Closed RangerMauve closed 8 months ago
Test are failing with NameError: name 'Field' is not defined
- I think there's a missing import.
Ty, added the imports from llama-cpp.
Tested out with llm install ./
which worked out. The options seem to be getting passed in.
Tweaking params didn't make the replit model perform well enough to be useful to me sadly :P
With this PR checked out, llm models --options
included:
gpt4all: all-MiniLM-L6-v2-f16 - SBert, 43.76MB download, needs 1GB RAM (installed)
max_tokens: int
The maximum number of tokens to generate.
temp: float
The model temperature. Larger values increase creativity but decrease
factuality.
top_k: int
Randomly sample from the top_k most likely tokens at each generation
step. Set this to 1 for greedy decoding.
top_p: float
Randomly sample at each generation step from the top most likely
tokens whose probabilities add up to top_p.
repeat_penalty: float
Penalize the model for repetition. Higher values result in less
repetition.
repeat_last_n: int
How far in the models generation history to apply the repeat penalty.
n_batch: int
Number of prompt tokens processed in parallel. Larger values decrease
latency but increase resource requirements.
gpt4all: orca-mini-3b-gguf2-q4_0 - Mini Orca (Small), 1.84GB download, needs 4GB RAM (installed)
max_tokens: int
temp: float
top_k: int
top_p: float
repeat_penalty: float
repeat_last_n: int
n_batch: int
gpt4all: mistral-7b-instruct-v0 - Mistral Instruct, 3.83GB download, needs 8GB RAM (installed)
max_tokens: int
temp: float
top_k: int
top_p: float
repeat_penalty: float
repeat_last_n: int
n_batch: int
$ llm -m mistral-7b-instruct-v0 'hello' -o max_tokens 2
Hello!
$ llm -m mistral-7b-instruct-v0 'hello' -o max_tokens 10
Hello! How can I help you today?
Refs:
3
Per our discussion on the fediverse.
I took the parameter descriptions from the gpt4all python docs and copied the structure from the llm-llama-cpp plugin
Does this work? Any other changes that I should make?