Lightning-AI/lit-llama
Implementation of the LLaMA language model based on nanoGPT. Supports flash attention, Int8 and GPTQ 4bit quantization, LoRA and LLaMA-Adapter fine-tuning, pre-training. Apache 2.0-licensed.
PythonApache-2.0
Pinned issues
Issues
- 1
Where is tokenizer.model? tokenizer path
#488 opened by andreamigliorati - 0
数据集与训练方法的相关问题
#489 opened by feizhuqwq - 6
Running into StopIteration with single node multi GPU pretraining against the redpajama sample
#465 opened by cabal-daniel - 5
Using llama3 through lit lama
#487 opened by fireyanci - 1
- 0
why cannot the generate function be used twice
#485 opened by WyGongya - 0
Converting from lit-llama to HF checkpoint?
#484 opened by jacqueline-he - 0
it seems that hash of traindata is lost, so it's impossible to continue finetune after stop
#483 opened by drazdra - 0
- 4
OSError: Not found: "checkpoints/lit-llama/tokenizer.model": No such file or directory Error #2
#481 opened by anirudhitagi - 0
- 0
TPU Training
#480 opened by kathir-ks - 1
Mistral Model
#458 opened by PierreColombo - 1
RuntimeError: Expected x1.dtype() == cos.dtype() to be true, but got false. (Could this error message be improved? If so, please report an enhancement request to PyTorch.)
#472 opened by JerryDaHeLian - 0
Issue with Rotary Embedding Initialization when the number of devices is > 1
#479 opened by diegodoimo - 0
Beam search generation
#477 opened by hellowoe23 - 1
- 0
Ban some tokens
#474 opened by AnnaKholkina - 0
- 1
RuntimeError: cutlassF: no kernel found to launch!
#471 opened by xvanQ - 0
How to quantize LLama in fine-tuning ?
#470 opened by sfarzi - 0
How to convert hf weight of 70b to lit-lamma weights?
#469 opened by sfarzi - 3
Looking for LLaMA 2?
#452 opened by carmocca - 2
Full fine-tuning on Alpaca dataset with 4 L40s GPUs fails 8 hours into the training job with index_copy_
#468 opened by cabal-daniel - 0
- 0
- 0
How to do conversation with fine tuned model?
#464 opened by Harsh-raj - 0
[question] nan loss value and run time error
#463 opened by nevermet - 0
[question] assert lora_path.is_file() error
#462 opened by nevermet - 2
[question] error message while finetuning
#460 opened by nevermet - 0
Question about 'validating...' from lora.py
#461 opened by nevermet - 0
- 0
When I finetuned the model, an error occurred during the decoding process: IndexError: Out of range: piece id is out of range.
#457 opened by HypherX - 3
- 1
lightning llama
#436 opened by sri9s - 2
How can I do to inferece with different promts in Juypter Notebook, just load the model and tokenizer once?
#450 opened by Vinter8848 - 0
Only add a linear layer to LLaMA without any computation degrade the performance
#451 opened by YUCHEN005 - 1
- 1
(documentation) How do I know if generate.py is running on GPU / GPU configuration
#449 opened by maathieu - 1
ValueError: Precision 'bf16-true' is invalid
#447 opened by AlphaGoMK - 3
multi gpus for full finetune
#446 opened by qiqiApink - 7
- 2
No response after training an epoch
#443 opened by Dylandtt - 2
- 1
LORA: RuntimeError: GET was unable to find an engine to execute this computation
#441 opened by LamOne1 - 0
converting Adapter to huggingface format
#440 opened by LamOne1 - 1
Support conversion to huggingface format
#439 opened by LamOne1 - 1
Getting AssertionError when saving a FSDP strat trained model with 16-mixed precision
#434 opened by JimenezBarreroDavid - 0
combine adapter weights with the base model
#433 opened by wxl-lxw - 0
Question about FlashAttention and KV-cache
#430 opened by KnowingNothing