wangqiang9 / sketchknitter Goto Github PK

The evaluator needs some instructions on how to use it.

I'm testing the evaluator but the read_activations function does not accept the reference dataset.
python evaluations/evaluator.py datasets/airplane.npz datasets/airplane.npz

The above command throws the below error:
ValueError: missing arr_0 in npz file

Can you also advise on the right format for samples? Its not very clear from the code.

When I checked code, I couldn't find .losses.
Please let me know detail.
gaussian_diffusion.py line 8 :
from .losses import normal_kl, discretized_gaussian_log_likelihood ()

TypeError: ddim_sample_loop() missing 3 required positional arguments: 'data', 'raster', and 'loss'

When I set N to a smaller number, I ran into the following problem

ValueError: could not broadcast input array from shape (59,3) into shape (24,3)’

I tried to use my modified code and the result was very bad
len_seq = min(self.Nmax, self.sketches_normed[idx].shape[0])
step = self.sketches_normed[idx].shape[0] // len_seq + 1
sampledata = self.sketches_normed[idx][::step]
sketch[:len(sampledata), :] = sampledata
I see that your paper mentions data about N=24, how did you achieve it？

I want to train on my own sketch RGB dataset, but I do not think the sketch-rnn repo has the guidance of how to do that.
Is there a way to transform RGB dataset into your dataset format?

Thank you for sharing your compelling and meticulously constructed work. I attempted to reproduce the results of SketchKnitter using only the airplane class from the QuickDraw dataset. However, the quality of the generated airplane sketches did not align with my expectations, and the quantitative results were not as favorable as I had anticipated. I employed the code and dataset from your GitHub repository, executing the following commands to train SketchKnitter:

python train.py --data_dir [/path/to/datasets] \
                --lr 1e-4 \
                --batch_size 512 \
                --use_fp16 False \
                --log_dir [/path/to/log] \
                --diffusion_steps 1000 \
                --noise_schedule linear \
                --image_size 96 \
                --num_channels 96 \
                --num_res_blocks 3

I also sampled the sketches and evaluated the quality using your code. The results are as follows:

• Inception Score: 2.460366725921631
• FID: 61.48920809233371
• sFID: 49.6528623255079
• Precision: 0.5416
• Recall: 0.37

I suspect there may be some issues with my approach. However, I haven't made any modifications to your original code. Could you please provide any guidance or suggestions on potential adjustments I could make to improve the performance? I appreciate your time and help in this matter. Have a wonderful day, and I look forward to your response.

I'm trying to use SketchKnitter to rectify bad sketches, but I can't find the corresponding code (including how to use the NLR module). Can you please upload the code for conditional generation?

Hi there!
Firstly, thank you for your advice about sample_all in sample.py! However, I noticed that in your dataset such as moon.npz, the size of data.train (as shown in the 80th line in draw_sketch.py) is 70000×3. While the size I get in sample_all.npz (from the 73rd line in sample.py) is 16×96×4.
So could you please explain how you save sample_all into the true npz file? I've tried
np.savez(args.save_path + "sample0.npz", train=sample_all) but it didn't work when draw_sketch.
Thank you for your help agian!

双向RNN的实现在哪里

ModuleNotFoundError: No module named 'sketch_diffusion.train_util'

The part for the sampling has:
while len(all_images) * args.batch_size < args.num_samples:
However, all_images is net updated which makes the code run infinitely.

I'm trying to reproduce the results presented in the paper, however using the configurations provided in the scripts (train.sh and sample.sh) I cannot archive the results. Here is an example of the results I obtained training using a single dataset:

They appear to be composed of dashes instead of a continuous lines.
Is it possible to prove the configurations used to train and sample to obtain the results from the paper?

Thank you very much for the code, I have some questions I would like to ask.
The sample.py does not save any samples anywhere.

sample_all = bin_pen(sample_all, args.pen_break)
print(f"sample all {sample_all} is saved!")

This code seems to be printing the whole vector ?

num_res_blocks=4,num_heads=8,my loss is around 0.08 and is no longer decreasing，and I'm wondering if I can stop training,but FID is 30,I used Apple in QuickDraw dataset,What can I do to lower my FID？

| grad_norm | 0.114 |
| loss | 0.0757 |
| loss_q0 | 0.227 |
| loss_q1 | 0.0749 |
| loss_q2 | 0.0142 |
| loss_q3 | 0.00181 |
| mse | 0.0747 |
| mse_q0 | 0.226 |
| mse_q1 | 0.074 |
| mse_q2 | 0.0132 |
| mse_q3 | 0.000843 |
| pen_state | 0.0965 |
| pen_state_q0 | 0.0964 |
| pen_state_q1 | 0.0966 |
| pen_state_q2 | 0.0966 |
| pen_state_q3 | 0.0965 |

Can I please get the model you're using in sample.py?
i'm getting this error
File "sample.py", line 45, in main
dist_util.load_state_dict(args.model_path, map_location="cpu")
File "/Users/filzahamjad/anaconda3/envs/sketchknitter/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1672, in load_state_dict
self.class.name, "\n\t".join(error_msgs)))
RuntimeError: Error(s) in loading state_dict for UNetModel:
Missing key(s) in state_dict: "time_embed.0.weight", "time_embed.0.bias", "time_embed.2.weight", "time_embed.2.bias", "input_blocks.0.0.weight", "input_blocks.0.0.bias", "input_blocks.1.0.in_layers.0.weight", "input_blocks.1.0.in_layers.0.bias", "input_blocks.1.0.in_layers.2.weight", "input_blocks.1.0.in_layers.2.bias", "input_blocks.1.0.emb_layers.1.weight", "input_blocks.1.0.emb_layers.1.bias", "input_blocks.1.0.out_layers.0.weight", "input_blocks.1.0.out_layers.0.bias", "input_blocks.1.0.out_layers.3.weight", "input_blocks.1.0.out_layers.3.bias", "input_blocks.2.0.in_layers.0.weight", "input_blocks.2.0.in_layers.0.bias", "input_blocks.2.0.in_layers.2.weight", "input_blocks.2.0.in_layers.2.bias", "input_blocks.2.0.emb_layers.1.weight", "input_blocks.2.0.emb_layers.1.bias", "input_blocks.2.0.out_layers.0.weight", "input_blocks.2.0.out_layers.0.bias", "input_blocks.2.0.out_layers.3.weight", "input_blocks.2.0.out_layers.3.bias", "input_blocks.3.0.in_layers.0.weight", "input_blocks.3.0.in_layers.0.bias", "input_blocks.3.0.in_layers.2.weight", "input_blocks.3.0.in_layers.2.bias", "input_blocks.3.0.emb_layers.1.weight", "input_blocks.3.0.emb_layers.1.bias", "input_blocks.3.0.out_layers.0.weight", "input_blocks.3.0.out_layers.0.bias", "input_blocks.3.0.out_layers.3.weight", "input_blocks.3.0.out_layers.3.bias", "input_blocks.4.0.op.weight", "input_blocks.4.0.op.bias", "input_blocks.5.0.in_layers.0.weight", "input_blocks.5.0.in_layers.0.bias", "input_blocks.5.0.in_layers.2.weight", "input_blocks.5.0.in_layers.2.bias", "input_blocks.5.0.emb_layers.1.weight", "input_blocks.5.0.emb_layers.1.bias", "input_blocks.5.0.out_layers.0.weight", "input_blocks.5.0.out_layers.0.bias", "input_blocks.5.0.out_layers.3.weight", "input_blocks.5.0.out_layers.3.bias", "input_blocks.5.0.skip_connection.weight", "input_blocks.5.0.skip_connection.bias", "input_blocks.6.0.in_layers.0.weight", "input_blocks.6.0.in_layers.0.bias", "input_blocks.6.0.in_layers.2.weight", "input_blocks.6.0.in_layers.2.bias", "input_blocks.6.0.emb_layers.1.weight", "input_blocks.6.0.emb_layers.1.bias", "input_blocks.6.0.out_layers.0.weight", "input_blocks.6.0.out_layers.0.bias", "input_blocks.6.0.out_layers.3.weight", "input_blocks.6.0.out_layers.3.bias", "input_blocks.7.0.in_layers.0.weight", "input_blocks.7.0.in_layers.0.bias", "input_blocks.7.0.in_layers.2.weight", "input_blocks.7.0.in_layers.2.bias", "input_blocks.7.0.emb_layers.1.weight", "input_blocks.7.0.emb_layers.1.bias", "input_blocks.7.0.out_layers.0.weight", "input_blocks.7.0.out_layers.0.bias", "input_blocks.7.0.out_layers.3.weight", "input_blocks.7.0.out_layers.3.bias", "input_blocks.8.0.op.weight", "input_blocks.8.0.op.bias", "input_blocks.9.0.in_layers.0.weight", "input_blocks.9.0.in_layers.0.bias", "input_blocks.9.0.in_layers.2.weight", "input_blocks.9.0.in_layers.2.bias", "input_blocks.9.0.emb_layers.1.weight", "input_blocks.9.0.emb_layers.1.bias", "input_blocks.9.0.out_layers.0.weight", "input_blocks.9.0.out_layers.0.bias", "input_blocks.9.0.out_layers.3.weight", "input_blocks.9.0.out_layers.3.bias", "input_blocks.9.0.skip_connection.weight", "input_blocks.9.0.skip_connection.bias", "input_blocks.10.0.in_layers.0.weight", "input_blocks.10.0.in_layers.0.bias", "input_blocks.10.0.in_layers.2.weight", "input_blocks.10.0.in_layers.2.bias", "input_blocks.10.0.emb_layers.1.weight", "input_blocks.10.0.emb_layers.1.bias", "input_blocks.10.0.out_layers.0.weight", "input_blocks.10.0.out_layers.0.bias", "input_blocks.10.0.out_layers.3.weight", "input_blocks.10.0.out_layers.3.bias", "input_blocks.11.0.in_layers.0.weight", "input_blocks.11.0.in_layers.0.bias", "input_blocks.11.0.in_layers.2.weight", "input_blocks.11.0.in_layers.2.bias", "input_blocks.11.0.emb_layers.1.weight", "input_blocks.11.0.emb_layers.1.bias", "input_blocks.11.0.out_layers.0.weight", "input_blocks.11.0.out_layers.0.bias", "input_blocks.11.0.out_layers.3.weight", "input_blocks.11.0.out_layers.3.bias", "input_blocks.12.0.op.weight", "input_blocks.12.0.op.bias", "input_blocks.13.0.in_layers.0.weight", "input_blocks.13.0.in_layers.0.bias", "input_blocks.13.0.in_layers.2.weight", "input_blocks.13.0.in_layers.2.bias", "input_blocks.13.0.emb_layers.1.weight", "input_blocks.13.0.emb_layers.1.bias", "input_blocks.13.0.out_layers.0.weight", "input_blocks.13.0.out_layers.0.bias", "input_blocks.13.0.out_layers.3.weight", "input_blocks.13.0.out_layers.3.bias", "input_blocks.13.0.skip_connection.weight", "input_blocks.13.0.skip_connection.bias", "input_blocks.14.0.in_layers.0.weight", "input_blocks.14.0.in_layers.0.bias", "input_blocks.14.0.in_layers.2.weight", "input_blocks.14.0.in_layers.2.bias", "input_blocks.14.0.emb_layers.1.weight", "input_blocks.14.0.emb_layers.1.bias", "input_blocks.14.0.out_layers.0.weight", "input_blocks.14.0.out_layers.0.bias", "input_blocks.14.0.out_layers.3.weight", "input_blocks.14.0.out_layers.3.bias", "input_blocks.15.0.in_layers.0.weight", "input_blocks.15.0.in_layers.0.bias", "input_blocks.15.0.in_layers.2.weight", "input_blocks.15.0.in_layers.2.bias", "input_blocks.15.0.emb_layers.1.weight", "input_blocks.15.0.emb_layers.1.bias", "input_blocks.15.0.out_layers.0.weight", "input_blocks.15.0.out_layers.0.bias", "input_blocks.15.0.out_layers.3.weight", "input_blocks.15.0.out_layers.3.bias", "middle_block.0.in_layers.0.weight", "middle_block.0.in_layers.0.bias", "middle_block.0.in_layers.2.weight", "middle_block.0.in_layers.2.bias", "middle_block.0.emb_layers.1.weight", "middle_block.0.emb_layers.1.bias", "middle_block.0.out_layers.0.weight", "middle_block.0.out_layers.0.bias", "middle_block.0.out_layers.3.weight", "middle_block.0.out_layers.3.bias", "middle_block.1.norm.weight", "middle_block.1.norm.bias", "middle_block.1.qkv.weight", "middle_block.1.qkv.bias", "middle_block.1.proj_out.weight", "middle_block.1.proj_out.bias", "middle_block.2.in_layers.0.weight", "middle_block.2.in_layers.0.bias", "middle_block.2.in_layers.2.weight", "middle_block.2.in_layers.2.bias", "middle_block.2.emb_layers.1.weight", "middle_block.2.emb_layers.1.bias", "middle_block.2.out_layers.0.weight", "middle_block.2.out_layers.0.bias", "middle_block.2.out_layers.3.weight", "middle_block.2.out_layers.3.bias", "output_blocks.0.0.in_layers.0.weight", "output_blocks.0.0.in_layers.0.bias", "output_blocks.0.0.in_layers.2.weight", "output_blocks.0.0.in_layers.2.bias", "output_blocks.0.0.emb_layers.1.weight", "output_blocks.0.0.emb_layers.1.bias", "output_blocks.0.0.out_layers.0.weight", "output_blocks.0.0.out_layers.0.bias", "output_blocks.0.0.out_layers.3.weight", "output_blocks.0.0.out_layers.3.bias", "output_blocks.0.0.skip_connection.weight", "output_blocks.0.0.skip_connection.bias", "output_blocks.1.0.in_layers.0.weight", "output_blocks.1.0.in_layers.0.bias", "output_blocks.1.0.in_layers.2.weight", "output_blocks.1.0.in_layers.2.bias", "output_blocks.1.0.emb_layers.1.weight", "output_blocks.1.0.emb_layers.1.bias", "output_blocks.1.0.out_layers.0.weight", "output_blocks.1.0.out_layers.0.bias", "output_blocks.1.0.out_layers.3.weight", "output_blocks.1.0.out_layers.3.bias", "output_blocks.1.0.skip_connection.weight", "output_blocks.1.0.skip_connection.bias", "output_blocks.2.0.in_layers.0.weight", "output_blocks.2.0.in_layers.0.bias", "output_blocks.2.0.in_layers.2.weight", "output_blocks.2.0.in_layers.2.bias", "output_blocks.2.0.emb_layers.1.weight", "output_blocks.2.0.emb_layers.1.bias", "output_blocks.2.0.out_layers.0.weight", "output_blocks.2.0.out_layers.0.bias", "output_blocks.2.0.out_layers.3.weight", "output_blocks.2.0.out_layers.3.bias", "output_blocks.2.0.skip_connection.weight", "output_blocks.2.0.skip_connection.bias", "output_blocks.3.0.in_layers.0.weight", "output_blocks.3.0.in_layers.0.bias", "output_blocks.3.0.in_layers.2.weight", "output_blocks.3.0.in_layers.2.bias", "output_blocks.3.0.emb_layers.1.weight", "output_blocks.3.0.emb_layers.1.bias", "output_blocks.3.0.out_layers.0.weight", "output_blocks.3.0.out_layers.0.bias", "output_blocks.3.0.out_layers.3.weight", "output_blocks.3.0.out_layers.3.bias", "output_blocks.3.0.skip_connection.weight", "output_blocks.3.0.skip_connection.bias", "output_blocks.3.1.conv.weight", "output_blocks.3.1.conv.bias", "output_blocks.4.0.in_layers.0.weight", "output_blocks.4.0.in_layers.0.bias", "output_blocks.4.0.in_layers.2.weight", "output_blocks.4.0.in_layers.2.bias", "output_blocks.4.0.emb_layers.1.weight", "output_blocks.4.0.emb_layers.1.bias", "output_blocks.4.0.out_layers.0.weight", "output_blocks.4.0.out_layers.0.bias", "output_blocks.4.0.out_layers.3.weight", "output_blocks.4.0.out_layers.3.bias", "output_blocks.4.0.skip_connection.weight", "output_blocks.4.0.skip_connection.bias", "output_blocks.5.0.in_layers.0.weight", "output_blocks.5.0.in_layers.0.bias", "output_blocks.5.0.in_layers.2.weight", "output_blocks.5.0.in_layers.2.bias", "output_blocks.5.0.emb_layers.1.weight", "output_blocks.5.0.emb_layers.1.bias", "output_blocks.5.0.out_layers.0.weight", "output_blocks.5.0.out_layers.0.bias", "output_blocks.5.0.out_layers.3.weight", "output_blocks.5.0.out_layers.3.bias", "output_blocks.5.0.skip_connection.weight", "output_blocks.5.0.skip_connection.bias", "output_blocks.6.0.in_layers.0.weight", "output_blocks.6.0.in_layers.0.bias", "output_blocks.6.0.in_layers.2.weight", "output_blocks.6.0.in_layers.2.bias", "output_blocks.6.0.emb_layers.1.weight", "output_blocks.6.0.emb_layers.1.bias", "output_blocks.6.0.out_layers.0.weight", "output_blocks.6.0.out_layers.0.bias", "output_blocks.6.0.out_layers.3.weight", "output_blocks.6.0.out_layers.3.bias", "output_blocks.6.0.skip_connection.weight", "output_blocks.6.0.skip_connection.bias", "output_blocks.7.0.in_layers.0.weight", "output_blocks.7.0.in_layers.0.bias", "output_blocks.7.0.in_layers.2.weight", "output_blocks.7.0.in_layers.2.bias", "output_blocks.7.0.emb_layers.1.weight", "output_blocks.7.0.emb_layers.1.bias", "output_blocks.7.0.out_layers.0.weight", "output_blocks.7.0.out_layers.0.bias", "output_blocks.7.0.out_layers.3.weight", "output_blocks.7.0.out_layers.3.bias", "output_blocks.7.0.skip_connection.weight", "output_blocks.7.0.skip_connection.bias", "output_blocks.7.1.conv.weight", "output_blocks.7.1.conv.bias", "output_blocks.8.0.in_layers.0.weight", "output_blocks.8.0.in_layers.0.bias", "output_blocks.8.0.in_layers.2.weight", "output_blocks.8.0.in_layers.2.bias", "output_blocks.8.0.emb_layers.1.weight", "output_blocks.8.0.emb_layers.1.bias", "output_blocks.8.0.out_layers.0.weight", "output_blocks.8.0.out_layers.0.bias", "output_blocks.8.0.out_layers.3.weight", "output_blocks.8.0.out_layers.3.bias", "output_blocks.8.0.skip_connection.weight", "output_blocks.8.0.skip_connection.bias", "output_blocks.9.0.in_layers.0.weight", "output_blocks.9.0.in_layers.0.bias", "output_blocks.9.0.in_layers.2.weight", "output_blocks.9.0.in_layers.2.bias", "output_blocks.9.0.emb_layers.1.weight", "output_blocks.9.0.emb_layers.1.bias", "output_blocks.9.0.out_layers.0.weight", "output_blocks.9.0.out_layers.0.bias", "output_blocks.9.0.out_layers.3.weight", "output_blocks.9.0.out_layers.3.bias", "output_blocks.9.0.skip_connection.weight", "output_blocks.9.0.skip_connection.bias", "output_blocks.10.0.in_layers.0.weight", "output_blocks.10.0.in_layers.0.bias", "output_blocks.10.0.in_layers.2.weight", "output_blocks.10.0.in_layers.2.bias", "output_blocks.10.0.emb_layers.1.weight", "output_blocks.10.0.emb_layers.1.bias", "output_blocks.10.0.out_layers.0.weight", "output_blocks.10.0.out_layers.0.bias", "output_blocks.10.0.out_layers.3.weight", "output_blocks.10.0.out_layers.3.bias", "output_blocks.10.0.skip_connection.weight", "output_blocks.10.0.skip_connection.bias", "output_blocks.11.0.in_layers.0.weight", "output_blocks.11.0.in_layers.0.bias", "output_blocks.11.0.in_layers.2.weight", "output_blocks.11.0.in_layers.2.bias", "output_blocks.11.0.emb_layers.1.weight", "output_blocks.11.0.emb_layers.1.bias", "output_blocks.11.0.out_layers.0.weight", "output_blocks.11.0.out_layers.0.bias", "output_blocks.11.0.out_layers.3.weight", "output_blocks.11.0.out_layers.3.bias", "output_blocks.11.0.skip_connection.weight", "output_blocks.11.0.skip_connection.bias", "output_blocks.11.1.conv.weight", "output_blocks.11.1.conv.bias", "output_blocks.12.0.in_layers.0.weight", "output_blocks.12.0.in_layers.0.bias", "output_blocks.12.0.in_layers.2.weight", "output_blocks.12.0.in_layers.2.bias", "output_blocks.12.0.emb_layers.1.weight", "output_blocks.12.0.emb_layers.1.bias", "output_blocks.12.0.out_layers.0.weight", "output_blocks.12.0.out_layers.0.bias", "output_blocks.12.0.out_layers.3.weight", "output_blocks.12.0.out_layers.3.bias", "output_blocks.12.0.skip_connection.weight", "output_blocks.12.0.skip_connection.bias", "output_blocks.13.0.in_layers.0.weight", "output_blocks.13.0.in_layers.0.bias", "output_blocks.13.0.in_layers.2.weight", "output_blocks.13.0.in_layers.2.bias", "output_blocks.13.0.emb_layers.1.weight", "output_blocks.13.0.emb_layers.1.bias", "output_blocks.13.0.out_layers.0.weight", "output_blocks.13.0.out_layers.0.bias", "output_blocks.13.0.out_layers.3.weight", "output_blocks.13.0.out_layers.3.bias", "output_blocks.13.0.skip_connection.weight", "output_blocks.13.0.skip_connection.bias", "output_blocks.14.0.in_layers.0.weight", "output_blocks.14.0.in_layers.0.bias", "output_blocks.14.0.in_layers.2.weight", "output_blocks.14.0.in_layers.2.bias", "output_blocks.14.0.emb_layers.1.weight", "output_blocks.14.0.emb_layers.1.bias", "output_blocks.14.0.out_layers.0.weight", "output_blocks.14.0.out_layers.0.bias", "output_blocks.14.0.out_layers.3.weight", "output_blocks.14.0.out_layers.3.bias", "output_blocks.14.0.skip_connection.weight", "output_blocks.14.0.skip_connection.bias", "output_blocks.15.0.in_layers.0.weight", "output_blocks.15.0.in_layers.0.bias", "output_blocks.15.0.in_layers.2.weight", "output_blocks.15.0.in_layers.2.bias", "output_blocks.15.0.emb_layers.1.weight", "output_blocks.15.0.emb_layers.1.bias", "output_blocks.15.0.out_layers.0.weight", "output_blocks.15.0.out_layers.0.bias", "output_blocks.15.0.out_layers.3.weight", "output_blocks.15.0.out_layers.3.bias", "output_blocks.15.0.skip_connection.weight", "output_blocks.15.0.skip_connection.bias", "out.0.weight", "out.0.bias", "out.2.weight", "out.2.bias", "pen_state_out.0.weight", "pen_state_out.0.bias".
Unexpected key(s) in state_dict: "state", "param_groups".

TypeError: ddim_sample_loop() missing 3 required positional arguments: 'data', 'raster', and 'loss'
I used below samping command.
python sample.py --model_path ./log/ema_0.9999_050000.pt --pen_break 0.1 --save_path ./ --use_ddim True --log_dir ./log/ --diffusion_steps 100 --noise_schedule linear --image_size 96 --num_channels 96 --num_res_blocks 3

Thank you so much for your excellent work!
I can't find how to save "sample_all" in .npz format and draw sketch from them. In sample.py, the main function stops at these two statements:
sample_all = th.cat((sample, pen_state), 2).cpu()
sample_all = bin_pen(sample_all, args.pen_break)
So could you please explain how to save "sample_all" in .npz format and get a sketch result ?