princetonvisualai / multimodal_dataset_distillation Goto Github PK

hi, when creating the env using conda, the following errors arise:
Collecting package metadata (repodata.json): | WARNING conda.models.version:get_matcher(556): Using .* with relational operator is superfluous and deprecated and will be removed in a future version of conda. Your spec was 1.9.0., but conda is ignoring the . and treating it as 1.9.0
WARNING conda.models.version:get_matcher(556): Using .* with relational operator is superfluous and deprecated and will be removed in a future version of conda. Your spec was 1.6.0., but conda is ignoring the . and treating it as 1.6.0
WARNING conda.models.version:get_matcher(556): Using .* with relational operator is superfluous and deprecated and will be removed in a future version of conda. Your spec was 1.8.0., but conda is ignoring the . and treating it as 1.8.0
WARNING conda.models.version:get_matcher(556): Using .* with relational operator is superfluous and deprecated and will be removed in a future version of conda. Your spec was 1.7.1., but conda is ignoring the . and treating it as 1.7.1

the file requirements.yaml is generated automatically?? Since it contains so many unecessary pacakges and is not a common installation format, will you make a regular requirements.yaml or requirements.txt for easier use, thank u.

Dear Author,
I noted that the new version of your paper in arxiv has "Low-Rank Adaptation Matching". I'm interested in how to implement it but I can't find the corresponding code in distill.py or buffer.py. Could you please share the code of Low-Rank Adaptation Matching.

In buffer.py, there is a line of code on line 39: data = np.load(f'{args.dataset}_{args.text_encoder}text_embed.
The error message is as follows: Traceback (most recent call last)
File "/root/autodl-tmp/multi/buffer.py", line 168, in
main(args)
File "/root/autodl-tmp/multi/buffer.py", line 39, in main
data = np.load(f'{args.dataset}{args.text_encoder}_text_embed.npz')
The File "/ root/miniconda3 / envs/distillation/lib/python3.9 / site - packages/numpy/lib/npyio. Py", line 417, in the load
fid = stack.enter_context(open(os_fspath(file), "rb"))
FileNotFoundError: [Errno 2] No such file or directory: 'flickr_bert_text_embed.npz'
Where can I download "flickr_bert_text_embed.npz"?

I used a syn_step=2&4 setting and ran the experiment on 8 V100 GPUs, but the results were very poor, with i2t and t2I rank1 only at 0.3. What could be the reason? The distillation loss is not decreasing. Should the hyperparameters be adjusted under this setting?