Search results

Remove jsonl.lock files in cache once they are no longer needed

Status: Open.

#74 In johny-b/llmcomp;

· johny-b opened on May 6, 2026

Maybe use hyperparameter_utils.get_lr as default learning rate?

Status: Open.

#72 In johny-b/llmcomp;

· Butanium opened on May 5, 2026

Config values should be contyext managers

Status: Open.

#71 In johny-b/llmcomp;

· johny-b opened on Apr 24, 2026

batch of openai fientuning jobs

Status: Open.

#70 In johny-b/llmcomp;

· Butanium opened on Apr 24, 2026

GPT-5.4 doesn't seem to support other temperatures than 0? Not sure, check.

bugSomething isn't workingSomething isn't working

Status: Open.

#69 In johny-b/llmcomp;

· johny-b opened on Apr 20, 2026

Implement default prompt caching

enhancementNew feature or requestNew feature or request

Status: Open.

#66 In johny-b/llmcomp;

· johny-b opened on Mar 25, 2026

Compress error messages

enhancementNew feature or requestNew feature or request

Status: Open.

#65 In johny-b/llmcomp;

· johny-b opened on Mar 18, 2026

Allow multiple paraphrases in RatingJudge

enhancementNew feature or requestNew feature or request

Status: Open.

#63 In johny-b/llmcomp;

· johny-b opened on Mar 11, 2026

Unexpected API error + logprobs-using Question = hard to trace error

Status: Open.

#62 In johny-b/llmcomp;

· johny-b opened on Mar 2, 2026

Reference examples in docstrings etc

Status: Open.

#57 In johny-b/llmcomp;

· johny-b opened on Feb 12, 2026

Simplify implementation around plots/views

Status: Open.

#56 In johny-b/llmcomp;

· johny-b opened on Feb 12, 2026

Models that were trained from a previous finetune

enhancementNew feature or requestNew feature or request

finetuning

Status: Open.

#54 In johny-b/llmcomp;

· johny-b opened on Feb 9, 2026