ilya16 commited on
Commit
cc610df
·
verified ·
1 Parent(s): a2499cb

Update README.md

Browse files

Add `library_name` and update example code for `symupe==1.1.0 `

Files changed (1) hide show
  1. README.md +16 -20
README.md CHANGED
@@ -1,4 +1,5 @@
1
  ---
 
2
  license: cc-by-nc-sa-4.0
3
  datasets:
4
  - SyMuPe/PERiScoPe
@@ -34,41 +35,36 @@ Introduced in the paper: [**SyMuPe: Affective and Controllable Symbolic Music Pe
34
 
35
  ## Quick Start
36
 
37
- To use this model, ensure you have the `symupe` library installed (refer to the [GitHub repo](https://github.com/ilya16/SyMuPe) for installation instructions).
 
 
 
 
 
38
 
39
  ```python
40
  import torch
41
  from symusic import Score
42
-
43
- from symupe.data.tokenizers import SyMuPe
44
- from symupe.inference import AutoGenerator, perform_score, save_performances
45
- from symupe.models import AutoModel
46
 
47
  device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
48
 
49
- # Load the model and tokenizer directly from the Hub
50
- model = AutoModel.from_pretrained("SyMuPe/MLM-base").to(device)
51
- tokenizer = SyMuPe.from_pretrained("SyMuPe/MLM-base")
52
-
53
- # Prepare generator for the model
54
- generator = AutoGenerator.from_model(model, tokenizer, device=device)
55
-
56
- # Load score MIDI
57
- score_midi = Score("score.mid")
58
 
59
  # Perform score MIDI (tokenization is handled inside)
60
- gen_results = perform_score(
61
- generator=generator,
62
- score=score_midi,
63
  use_score_context=True,
64
  num_samples=8,
65
- seed=23
66
  )
67
  # gen_results[i] is PerformanceRenderingResult(...) containing:
68
  # - score_midi, score_seq, gen_seq, perf_seq, perf_midi, perf_midi_sus
69
 
70
- # Save performed MIDI files in a single directory
71
- save_performances(gen_results, out_dir="samples/mlm", save_midi=True)
72
  ```
73
 
74
  ## License
 
1
  ---
2
+ library_name: symupe
3
  license: cc-by-nc-sa-4.0
4
  datasets:
5
  - SyMuPe/PERiScoPe
 
35
 
36
  ## Quick Start
37
 
38
+ Before using this model, ensure you have the `symupe` library installed:
39
+ ```shell
40
+ pip install -U symupe
41
+ ```
42
+
43
+ Use the following code to render performances:
44
 
45
  ```python
46
  import torch
47
  from symusic import Score
48
+ from symupe import AutoGenerator
 
 
 
49
 
50
  device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
51
 
52
+ # Build Generator by loading the model and tokenizer directly from the Hub
53
+ generator = AutoGenerator.from_pretrained("SyMuPe/MLM-base", device=device)
54
+ # model, tokenizer = generator.model, generator.tokenizer
 
 
 
 
 
 
55
 
56
  # Perform score MIDI (tokenization is handled inside)
57
+ gen_results = generator.perform_score(
58
+ "score.mid",
 
59
  use_score_context=True,
60
  num_samples=8,
61
+ seed=23,
62
  )
63
  # gen_results[i] is PerformanceRenderingResult(...) containing:
64
  # - score_midi, score_seq, gen_seq, perf_seq, perf_midi, perf_midi_sus
65
 
66
+ # Save performed MIDI files
67
+ generator.save_performances(gen_results, out_dir="samples/mlm")
68
  ```
69
 
70
  ## License