philschmid/sagemaker-huggingface-llama-2-samples

Issues

• The model did not return a loss from the inputs

  #22 opened a year ago by EtienneFerrandi
  0

• No matching distribution found for peft==0.4.0

  #21 opened a year ago by EtienneFerrandi
  4

• ValueError: expected sequence of length 2048 at dim 1 (got 0)

  #20 opened a year ago by vipinap98
  0

• OSError: meta-llama/Llama-2-13b-hf is not a local folder and is not a valid model identifier listed on 'https://huggingface.co/models'

  #6 opened a year ago by sashakttripathi
  2

• HF token leak

  #19 opened a year ago by tcapelle
  1

• Does we need the prompt formatting during the instruct model fine-tuning?

  #18 opened a year ago by mariababich
  0

• Error running fine-tuning on SageMaker

  #17 opened a year ago by abhyudit309
  4

• Fine Tuning LLAMA2-13b on ml.g4dn.12xlarge is taking too much time.

  #15 opened a year ago by hz-nm
  5

• Does the inference notebook work with Code Llama models?

  #16 opened a year ago by jackkwok
  4

• Error in build_llama2_prompt()

  #13 opened a year ago by pleonova
  1

• pretrain llama-2

  #12 opened a year ago by enpassanty
  0

• Cannot save checkpoints on provided s3 bucket

  #10 opened a year ago by AnnamariaGalante
  4

• Confused about chunk and potential concern of it hurting finetuning performance

  #11 opened a year ago by kalvin1024
  0

• Error running trained model on sagemaker endpoint

  #9 opened a year ago by sdkramer10
  3

• RuntimeError: cannot reshape tensor

  #8 opened a year ago by mikewadhera
  1

• Inference generating output with <unks> only with batch decoding but generates correct result if I pass a single prompt.

  #5 opened a year ago by irshadbhat
  2

• trainer.evaluate() with run_clm.py

  #7 opened a year ago by sadransh
  0

• missing "requirements.txt"

  #1 opened a year ago by andreaolgiati
  5

• No models to quantize

  #2 opened a year ago by andreaolgiati
  2

• CUDA error encountered on both BS=2&3 for 7B and 13B.

  #4 opened a year ago by alfredcs
  1

• RuntimeError: mat1 and mat2 shapes cannot be multiplied (4096x5120 and 1x2560)

  #3 opened a year ago by monuminu
  1