Issue with convert 7B weights in ubuntu 20.04 LTS

[kaleavess]

Hi All

I am very new using AI model and I tried a few of translation models like Aplaca.cpp & GPT4AALL however those are base on 7B. I want to run a 30B/65B in my server. I followed the installation guide start with 7B for first trial its without problem till convert the .pth file to .ggml.

Error massage showed below:
llama@llama:~/llama.cpp$ python3 convert.py models/7B/
Loading model file models/7B/consolidated.00.pth
Loading vocab file models/tokenizer.model
Writing vocab...
[ 1/291] Writing tensor tok_embeddings.weight | size 32000 x 4096 | type UnquantizedDataType(name='F16')
[ 2/291] Writing tensor norm.weight | size 4096 | type UnquantizedDataType(name='F32')
[ 3/291] Writing tensor output.weight | size 32000 x 4096 | type UnquantizedDataType(name='F16')
Traceback (most recent call last):
File "/home/llama/llama.cpp/convert.py", line 1149, in <module>
main()
File "/home/llama/llama.cpp/convert.py", line 1144, in main
OutputFile.write_all(outfile, params, model, vocab)
File "/home/llama/llama.cpp/convert.py", line 953, in write_all
for i, ((name, lazy_tensor), ndarray) in enumerate(zip(model.items(), ndarrays)):
File "/home/llama/llama.cpp/convert.py", line 875, in bounded_parallel_map
result = futures.pop(0).result()
File "/usr/lib/python3.10/concurrent/futures/_base.py", line 451, in result
return self.__get_result()
File "/usr/lib/python3.10/concurrent/futures/_base.py", line 403, in __get_result
raise self._exception
File "/usr/lib/python3.10/concurrent/futures/thread.py", line 58, in run
result = self.fn(*self.args, **self.kwargs)
File "/home/llama/llama.cpp/convert.py", line 950, in do_itemv
return lazy_tensor.load().to_ggml().ndarray
File "/home/llama/llama.cpp/convert.py", line 489, in load
ret = self._load()
File "/home/llama/llama.cpp/convert.py", line 497, in load
return self.load().astype(data_type)
File "/home/llama/llama.cpp/convert.py", line 489, in load
ret = self._load()
File "/home/llama/llama.cpp/convert.py", line 695, in load
return UnquantizedTensor(storage.load(storage_offset, elm_count).reshape(size))
File "/home/llama/llama.cpp/convert.py", line 680, in load
fp = self.zip_file.open(info)
File "/usr/lib/python3.10/zipfile.py", line 1535, in open
raise BadZipFile("Bad magic number for file header")
zipfile.BadZipFile: Bad magic number for file header

Machine spec:
CPU: Ryzen 5700G
GPU: RTX 2060 it 12GB
Ram: 32GB

Thank you so much

[Azeirah]

Are you certain the model you downloaded is not corrupt? The error you're getting is that one of the zipfiles is missing a header that all zipfiles should have. Is your download of the 7B model unfinished perhaps? Not sure

[prusnak]

Check downloaded files via sha256sum --ignore-missing -c SHA256SUMS

Please reopen if downloaded files are giving correct hashes.