GGUF Creation from Llamafy
#1
by
RonanMcGovern
- opened
Thanks for making this. I had hoped this mix allow the model to be GGUF'd but an error persists:
params = Params(n_vocab=151936, n_embd=4096, n_layer=32, n_ctx=32768, n_ff=11008, n_head=32, n_head_kv=32, n_experts=None, n_experts_used=None, f_norm_eps=1e-06, rope_scaling_type=None, f_rope_freq_base=1000000.0, f_rope_scale=None, n_orig_ctx=None, rope_finetuned=None, ftype=None, path_model=PosixPath('models'))
Found vocab files: {'tokenizer.model': None, 'vocab.json': PosixPath('models/vocab.json'), 'tokenizer.json': PosixPath('models/tokenizer.json')}
Loading vocab file 'models/vocab.json', type 'spm'
Traceback (most recent call last):
File "/workspace/llama.cpp/convert.py", line 1483, in <module>
main()
File "/workspace/llama.cpp/convert.py", line 1451, in main
vocab, special_vocab = vocab_factory.load_vocab(args.vocab_type, model_parent_path)
File "/workspace/llama.cpp/convert.py", line 1336, in load_vocab
vocab = SentencePieceVocab(
File "/workspace/llama.cpp/convert.py", line 394, in __init__
self.sentencepiece_tokenizer = SentencePieceProcessor(str(fname_tokenizer))
File "/usr/local/lib/python3.10/dist-packages/sentencepiece/__init__.py", line 447, in Init
self.Load(model_file=model_file, model_proto=model_proto)
File "/usr/local/lib/python3.10/dist-packages/sentencepiece/__init__.py", line 905, in Load
return self.LoadFromFile(model_file)
File "/usr/local/lib/python3.10/dist-packages/sentencepiece/__init__.py", line 310, in LoadFromFile
return _sentencepiece.SentencePieceProcessor_LoadFromFile(self, arg)
RuntimeError: Internal: src/sentencepiece_processor.cc(1101) [model_proto->ParseFromArray(serialized.data(), serialized.size())]
After calling:
!python convert.py models/
I know your repo isn't about GGUF, but given you understand how to Llamafy, I thought you might have ideas?
tokenizer_class": "Qwen2Tokenizer",
tokenizer_class": "Qwen2Tokenizer",
Right, that's what is in this tokenizer_config file..
Are you saying that's the problem, and it can't be solved?
OR, are you saying to change that somehow?
Thanks.
Can you provide transformers version?
Sure:
- `transformers` version: 4.38.1
- Platform: Linux-5.4.0-139-generic-x86_64-with-glibc2.35
- Python version: 3.10.12
- Huggingface_hub version: 0.20.3
- Safetensors version: 0.4.2
- Accelerate version: 0.27.2
- Accelerate config: not found
- PyTorch version (GPU?): 2.1.1+cu121 (True)
- Tensorflow version (GPU?): not installed (NA)
- Flax version (CPU?/GPU?/TPU?): not installed (NA)
- Jax version: not installed
- JaxLib version: not installed
- Using GPU in script?: A6000
- Using distributed or parallel set-up in script?: Single GPU
I don't know.... Maybe you can download tokenizer.model again
Ok, no problem, I appreciate the help.
BTW, I don't see tokenizer.model, do you mean tokenizer_config.json?