convert-hf : print output file name when completed

[danbev]

This commit adds the output file name to the log message when the conversion is completed.

The motivation for this change is that when --outfile option is not specified it migth not be obvious where the output file is written.

With this change the output of running the script will be something like the following:

INFO:hf-to-gguf:Model successfully exported to models/gemma-2-9b-it.gguf.

• I have read the contributing guidelines
• Self-reported review complexity:
  • Low
  • Medium
  • High

[compilade]

This is nice, although since #6942, the output file names are already printed before the progress bar.

On master, after printing the tensors, their types and their shape, there's this:

INFO:gguf.gguf_writer:Writing the following files:
INFO:gguf.gguf_writer:/srv/LLMstash/tmp/bf16-tests/tinymistral-bf16.gguf: n_tensors = 111, total_size = 496.1M
Writing: 100%|█████████████████████████████████████████████████████████████████████████████████| 496M/496M [00:05<00:00, 94.4Mbyte/s]
INFO:hf-to-gguf:Model successfully exported.

Your proposal would change this into

INFO:gguf.gguf_writer:Writing the following files:
INFO:gguf.gguf_writer:/srv/LLMstash/tmp/bf16-tests/tinymistral-bf16.gguf: n_tensors = 111, total_size = 496.1M
Writing: 100%|█████████████████████████████████████████████████████████████████████████████████| 496M/496M [00:05<00:00, 94.4Mbyte/s]
INFO:hf-to-gguf:Model successfully exported to /srv/LLMstash/tmp/bf16-tests/tinymistral-bf16.gguf.

Where the output file name is present twice.

And when using split options, the final name would not exist.

INFO:gguf.gguf_writer:Writing the following files:
INFO:gguf.gguf_writer:/srv/LLMstash/tmp/gguf-split-tests/tinymistral-100M/tinymistral-q8_0-00001-of-00003.gguf: n_tensors = 17, total_size = 99.5M
INFO:gguf.gguf_writer:/srv/LLMstash/tmp/gguf-split-tests/tinymistral-100M/tinymistral-q8_0-00002-of-00003.gguf: n_tensors = 58, total_size = 99.5M
INFO:gguf.gguf_writer:/srv/LLMstash/tmp/gguf-split-tests/tinymistral-100M/tinymistral-q8_0-00003-of-00003.gguf: n_tensors = 36, total_size = 64.7M
Shard (3/3): 100%|███████████████████████████████████████████████████████████████████████████| 64.7M/64.7M [00:00<00:00, 70.0Mbyte/s]
Writing: 100%|█████████████████████████████████████████████████████████████████████████████████| 264M/264M [00:07<00:00, 33.4Mbyte/s]
INFO:hf-to-gguf:Model successfully exported to /srv/LLMstash/tmp/gguf-split-tests/tinymistral-100M/tinymistral-q8_0.gguf.
$ ls /srv/LLMstash/tmp/gguf-split-tests/tinymistral-100M/
tinymistral-q8_0-00001-of-00003.gguf  tinymistral-q8_0-00002-of-00003.gguf  tinymistral-q8_0-00003-of-00003.gguf

But it might still be useful when not splitting.

[danbev]

This is nice, although since #6942, the output file names are already printed before the progress bar.

Ah sorry, I see it now, for some reason I just missed it :(

Now that I know where to look I'm happy with not having the output file name/path in the final output, but perhaps it might be nice to have it just the same in case others run into the same issue I did. I'll push a commit with a suggestion with the changes you suggested and also include printing the directory when the conversion performs a split. Thanks!

[compilade]

Ah sorry, I see it now, for some reason I just missed it :(

but perhaps it might be nice to have it just the same in case others run into the same issue I did.

I agree, this should reduce confusion.

Good idea with the output directory name when splitting.

[compilade]

Looks good to me.