gladjoyhub

How small should loss reach, 0.1 or 1.0?

LoRA training really slow

9

around 1 token per second, each evaluation take 30 minutes There are long lines (2000 words) in custom train.jsonl Please help, give a clue. Thanks! log: Iter 1: Val loss...

What's a good data format for lora fine-tuning?

4

I know the recommended format is this: {"text": "Q:What is the capital of France?\nA:The capital of France is Paris."} But some base model like Solar 10.4B recommends: .### User: What's...

incremental-indexeddb-adapter lokiid($loki) should start with 1 not 0

2

src/incremental-indexeddb-adapter.js line 69 to 72 ``` // chunkId - index of the data chunk - e.g. chunk 0 will be lokiIds 0-99 IncrementalIndexedDBAdapter.prototype._getChunk = function(collection, chunkId) { // 0-99, 100-199,...