Cannot build monsterapi/gemma-2b-lora-maths-orca-200k using LLM Conversion script

[hassanabidpk]

I am trying to convert monsterapi/gemma-2b-lora-maths-orca-200k to TFLite Flatbuffers file as mentioned here

But getting errors. I am reusing this script

Added below functions

def gemma2b_lora_download(token):
  REPO_ID = "monsterapi/gemma-2b-lora-maths-orca-200k"
  FILENAMES = ["tokenizer.json", "tokenizer_config.json", "adapter_model.safetensors"]
  os.environ['HF_TOKEN'] = "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"
  with out:
    for filename in FILENAMES:
      hf_hub_download(repo_id=REPO_ID, filename=filename, local_dir="./gemma-2b-lora-it")

def gemma2b_lora_convert_config(backend):
  input_ckpt = '/content/gemma-2b-lora-it/'
  vocab_model_file = '/content/gemma-2b-lora-it/'
  output_dir = '/content/intermediate/gemma-2b-lora-it/'
  output_tflite_file = f'/content/converted_models/gemma_base_lora_{backend}.bin'
  lora_output_tflite_file = f'/content/converted_models/gemma_lora_{backend}.bin'
  return converter.ConversionConfig(input_ckpt=input_ckpt, ckpt_format='safetensors', model_type='GEMMA_2B', backend=backend, output_dir=output_dir, combine_file_only=False, vocab_model_file=vocab_model_file, output_tflite_file=output_tflite_file, lora_output_tflite_file=lora_output_tflite_file)

Every time Colab enteprise is crashing

Is there any example to convert this trained model to .bin format for running on Android?

[kalyan2789g]

Hi @hassanabidpk,

We are looking into the issue. Could you please confirm the gemma model name/configuration which you are working? Please share the complete error logs to analyse the issue in detail.

We will keep you posted on the progress.

Thanks.

[hassanabidpk]

I am using this Gemma2B trained model : https://huggingface.co/monsterapi/gemma-2b-lora-maths-orca-200k based on google/gemma-2b

I am attaching
app.log
llm_conversion_notebook.txt
notebook and app.log file