sonithoiningombam commited on
Commit
29be0d7
·
verified ·
1 Parent(s): 3efd094

Add files using upload-large-folder tool

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .cache/Microsoft/DeveloperTools/deviceid +1 -0
  2. .conda/environments.txt +1 -0
  3. .gitattributes +3 -0
  4. .idea/this_studio.iml +56 -0
  5. .ipython/profile_default/db/dhist +0 -0
  6. .ipython/profile_default/startup/README +11 -0
  7. .lightning_studio/.studiorc +4 -0
  8. .lightning_studio/on_start.sh +15 -0
  9. .lightning_studio/on_stop.sh +10 -0
  10. .vscode/settings.json +90 -0
  11. bart/data/mni_train_tokenise.txt +3 -0
  12. data-bin/dict.eng_Latn.txt +0 -0
  13. data-bin/dict.mni_Beng.txt +0 -0
  14. data-bin/preprocess.log +14 -0
  15. data-bin/test.eng_Latn-mni_Beng.eng_Latn.idx +0 -0
  16. data-bin/test.eng_Latn-mni_Beng.mni_Beng.idx +0 -0
  17. data-bin/valid.eng_Latn-mni_Beng.eng_Latn.idx +0 -0
  18. data-bin/valid.eng_Latn-mni_Beng.mni_Beng.idx +0 -0
  19. data/checkpoints/checkpoint20.pt +3 -0
  20. data/fairseq/.github/CODEOWNERS +21 -0
  21. data/fairseq/.github/ISSUE_TEMPLATE.md +3 -0
  22. data/fairseq/.github/ISSUE_TEMPLATE/bug_report.md +43 -0
  23. data/fairseq/.github/ISSUE_TEMPLATE/documentation.md +15 -0
  24. data/fairseq/.github/ISSUE_TEMPLATE/feature_request.md +24 -0
  25. data/fairseq/.github/ISSUE_TEMPLATE/how-to-question.md +33 -0
  26. data/fairseq/.github/PULL_REQUEST_TEMPLATE.md +16 -0
  27. data/fairseq/.github/stale.yml +30 -0
  28. data/fairseq/.github/workflows/build.yml +81 -0
  29. data/fairseq/.github/workflows/depreview.yml +14 -0
  30. data/fairseq/.github/workflows/release.yml +161 -0
  31. data/fairseq/fairseq/clib/cuda/ngram_repeat_block_cuda_kernel.cu +82 -0
  32. data/fairseq/fairseq/clib/libbase/balanced_assignment.cpp +109 -0
  33. data/fairseq/fairseq/clib/libnat/edit_dist.cpp +231 -0
  34. data/fairseq/fairseq/distributed/__pycache__/__init__.cpython-38.pyc +0 -0
  35. data/fairseq/fairseq/distributed/__pycache__/fully_sharded_data_parallel.cpython-38.pyc +0 -0
  36. data/fairseq/fairseq/distributed/__pycache__/module_proxy_wrapper.cpython-38.pyc +0 -0
  37. data/fairseq/fairseq/distributed/__pycache__/tpu_distributed_data_parallel.cpython-38.pyc +0 -0
  38. data/fairseq/fairseq/distributed/__pycache__/utils.cpython-38.pyc +0 -0
  39. data/fairseq/fairseq/optim/__init__.py +48 -0
  40. data/fairseq/fairseq/optim/adadelta.py +47 -0
  41. data/fairseq/fairseq/optim/adam.py +239 -0
  42. data/fairseq/fairseq/optim/bmuf.py +200 -0
  43. data/train.eng_Latn +3 -0
  44. data/train.mni_Beng +3 -0
  45. nllb-mni-lsftl/checkpoint-26112/README.md +202 -0
  46. nllb-mni-lsftl/checkpoint-26112/adapter_config.json +31 -0
  47. nllb-mni-lsftl/checkpoint-26112/special_tokens_map.json +255 -0
  48. nllb-mni-lsftl/checkpoint-26112/tokenizer_config.json +1878 -0
  49. nllb-mni-lsftl/checkpoint-26112/trainer_state.json +0 -0
  50. nltk_data/corpora/omw-1.4.zip +3 -0
.cache/Microsoft/DeveloperTools/deviceid ADDED
@@ -0,0 +1 @@
 
 
1
+ 10c8f6e8-bcc8-4e89-8faf-6f1cba7a4d59
.conda/environments.txt ADDED
@@ -0,0 +1 @@
 
 
1
+ /home/zeus/miniconda3/envs/cloudspace
.gitattributes CHANGED
@@ -50,3 +50,6 @@ data/fairseq/fairseq/data/token_block_utils_fast.cpython-38-x86_64-linux-gnu.so
50
  nllb-mni-lsftl/checkpoint-26112/tokenizer.json filter=lfs diff=lfs merge=lfs -text
51
  data-bin/train.eng_Latn-mni_Beng.mni_Beng.idx filter=lfs diff=lfs merge=lfs -text
52
  data-bin/train.eng_Latn-mni_Beng.eng_Latn.idx filter=lfs diff=lfs merge=lfs -text
 
 
 
 
50
  nllb-mni-lsftl/checkpoint-26112/tokenizer.json filter=lfs diff=lfs merge=lfs -text
51
  data-bin/train.eng_Latn-mni_Beng.mni_Beng.idx filter=lfs diff=lfs merge=lfs -text
52
  data-bin/train.eng_Latn-mni_Beng.eng_Latn.idx filter=lfs diff=lfs merge=lfs -text
53
+ bart/data/mni_train_tokenise.txt filter=lfs diff=lfs merge=lfs -text
54
+ data/train.mni_Beng filter=lfs diff=lfs merge=lfs -text
55
+ data/train.eng_Latn filter=lfs diff=lfs merge=lfs -text
.idea/this_studio.iml ADDED
@@ -0,0 +1,56 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" encoding="UTF-8"?>
2
+ <module version="4">
3
+ <component name="NewModuleRootManager">
4
+ <content url="file://$MODULE_DIR$">
5
+ <excludeFolder url="file://$MODULE_DIR$/venv" />
6
+ <excludeFolder url="file://$MODULE_DIR$/lightning-annotations.json" />
7
+ <excludeFolder url="file://$MODULE_DIR$/.lightning-app-sync" />
8
+ <excludeFolder url="file://$MODULE_DIR$/.lightning-app-run" />
9
+ <excludeFolder url="file://$MODULE_DIR$/.lightning-upload" />
10
+ <excludeFolder url="file://$MODULE_DIR$/.ssh" />
11
+ <excludeFolder url="file://$MODULE_DIR$/.conda" />
12
+ <excludeFolder url="file://$MODULE_DIR$/.config" />
13
+ <excludeFolder url="file://$MODULE_DIR$/.npm" />
14
+ <excludeFolder url="file://$MODULE_DIR$/.nvm" />
15
+ <excludeFolder url="file://$MODULE_DIR$/.bower" />
16
+ <excludeFolder url="file://$MODULE_DIR$/.ipython" />
17
+ <excludeFolder url="file://$MODULE_DIR$/.local" />
18
+ <excludeFolder url="file://$MODULE_DIR$/.oh-my-zsh" />
19
+ <excludeFolder url="file://$MODULE_DIR$/.cache" />
20
+ <excludeFolder url="file://$MODULE_DIR$/miniconda3" />
21
+ <excludeFolder url="file://$MODULE_DIR$/.condarc" />
22
+ <excludeFolder url="file://$MODULE_DIR$/.gitconfig" />
23
+ <excludeFolder url="file://$MODULE_DIR$/.hushlogin" />
24
+ <excludeFolder url="file://$MODULE_DIR$/.profile" />
25
+ <excludeFolder url="file://$MODULE_DIR$/.screenrc" />
26
+ <excludeFolder url="file://$MODULE_DIR$/.vimrc" />
27
+ <excludeFolder url="file://$MODULE_DIR$/.sudo_as_admin_successful" />
28
+ <excludeFolder url="file://$MODULE_DIR$/.zcompdump" />
29
+ <excludeFolder url="file://$MODULE_DIR$/.bash_history" />
30
+ <excludeFolder url="file://$MODULE_DIR$/.bashrc" />
31
+ <excludeFolder url="file://$MODULE_DIR$/.zsh_history" />
32
+ <excludeFolder url="file://$MODULE_DIR$/.zshrc" />
33
+ <excludeFolder url="file://$MODULE_DIR$/.zshenv" />
34
+ <excludeFolder url="file://$MODULE_DIR$/.zlogin" />
35
+ <excludeFolder url="file://$MODULE_DIR$/.zprofile" />
36
+ <excludeFolder url="file://$MODULE_DIR$/.zlogout" />
37
+ <excludeFolder url="file://$MODULE_DIR$/.python_history" />
38
+ <excludeFolder url="file://$MODULE_DIR$/.lightningignore" />
39
+ <excludeFolder url="file://$MODULE_DIR$/.nv" />
40
+ <excludeFolder url="file://$MODULE_DIR$/.docker" />
41
+ <excludeFolder url="file://$MODULE_DIR$/.jupyter" />
42
+ <excludeFolder url="file://$MODULE_DIR$/.lightning" />
43
+ <excludeFolder url="file://$MODULE_DIR$/.vscode" />
44
+ <excludeFolder url="file://$MODULE_DIR$/.wget-hsts" />
45
+ <excludeFolder url="file://$MODULE_DIR$/.vscode-server" />
46
+ <excludeFolder url="file://$MODULE_DIR$/.vscode-server-insiders" />
47
+ <excludeFolder url="file://$MODULE_DIR$/.tmplaigit" />
48
+ <excludeFolder url="file://$MODULE_DIR$/.viminfo" />
49
+ <excludeFolder url="file://$MODULE_DIR$/.idea" />
50
+ <excludeFolder url="file://$MODULE_DIR$/.java" />
51
+ <excludeFolder url="file://$MODULE_DIR$/.dotnet" />
52
+ </content>
53
+ <orderEntry type="inheritedJdk" />
54
+ <orderEntry type="sourceFolder" forTests="false" />
55
+ </component>
56
+ </module>
.ipython/profile_default/db/dhist ADDED
Binary file (109 Bytes). View file
 
.ipython/profile_default/startup/README ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ This is the IPython startup directory
2
+
3
+ .py and .ipy files in this directory will be run *prior* to any code or files specified
4
+ via the exec_lines or exec_files configurables whenever you load this profile.
5
+
6
+ Files will be run in lexicographical order, so you can control the execution order of files
7
+ with a prefix, e.g.::
8
+
9
+ 00-first.py
10
+ 50-middle.py
11
+ 99-last.ipy
.lightning_studio/.studiorc ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ # This script is only for your user and runs in every shell you open.
2
+ # Use it to personalize your shell.
3
+ #
4
+ # Example: export MY_KEY=abcd-1234
.lightning_studio/on_start.sh ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/bin/bash
2
+
3
+ # This script runs every time your Studio starts, from your home directory.
4
+
5
+ # Logs from previous runs can be found in ~/.lightning_studio/logs/
6
+
7
+ # List files under fast_load that need to load quickly on start (e.g. model checkpoints).
8
+ #
9
+ # ! fast_load
10
+ # <your file here>
11
+
12
+ # Add your startup commands below.
13
+ #
14
+ # Example: streamlit run my_app.py
15
+ # Example: gradio my_app.py
.lightning_studio/on_stop.sh ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/bin/bash
2
+
3
+ # This script runs every time your Studio sleeps, from your home directory.
4
+
5
+ # Logs from previous runs can be found in ~/.lightning_studio/logs/
6
+
7
+ # Add your shutdown commands below.
8
+ #
9
+ # Example: docker down my-container
10
+ # Example: sudo service mysql stop
.vscode/settings.json ADDED
@@ -0,0 +1,90 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "python.defaultInterpreterPath": "/home/zeus/miniconda3/envs/cloudspace/bin/python",
3
+ "jupyter.kernels.excludePythonEnvironments": [
4
+ "/commands/python3",
5
+ "/commands/python",
6
+ "/bin/python3.10",
7
+ "/usr/bin/python3.10",
8
+ "/bin/python3",
9
+ "/usr/bin/python3",
10
+ "/bin/python3.8",
11
+ "/usr/bin/python3.8",
12
+ "/home/zeus/miniconda3/bin/python"
13
+ ],
14
+ "workbench.startupEditor": "none",
15
+ "python.terminal.activateEnvironment": false,
16
+ "terminal.integrated.drawBoldTextInBrightColors": false,
17
+ "terminal.integrated.gpuAcceleration": "on",
18
+ "terminal.integrated.localEchoLatencyThreshold": 0,
19
+ "terminal.integrated.localEchoEnabled": "off",
20
+ "terminal.integrated.localEchoStyle": "#000000",
21
+ "remote.autoForwardPorts": false,
22
+ "jupyter.notebookFileRoot": "${fileDirname}",
23
+ "terminal.integrated.enableMultiLinePasteWarning": "never",
24
+ "terminal.integrated.showTerminalConfigPrompt": false,
25
+ "files.autoSave": "afterDelay",
26
+ "files.exclude": {
27
+ "venv": true,
28
+ "lightning-annotations.json": true,
29
+ ".lightning-app-sync": true,
30
+ ".lighting-app-sync": true,
31
+ ".lightning-app-run": true,
32
+ "**/*.lightning_upload": true,
33
+ ".ssh*/**": true,
34
+ ".conda*/**": true,
35
+ ".config*/**": true,
36
+ ".npm*/**": true,
37
+ ".nvm*/**": true,
38
+ ".bower*/**": true,
39
+ ".ipython": true,
40
+ ".local*/**": true,
41
+ ".oh-my-zsh*/**": true,
42
+ "**/.cache/**": true,
43
+ "miniconda3": true,
44
+ ".condarc": true,
45
+ ".gitconfig": true,
46
+ ".hushlogin": true,
47
+ ".profile": true,
48
+ ".screenrc": true,
49
+ ".vimrc": true,
50
+ ".sudo_as_admin_successful": true,
51
+ ".zcompdump*": true,
52
+ ".bash_history": true,
53
+ ".bashrc": true,
54
+ ".zsh_history": true,
55
+ ".zshrc": true,
56
+ ".zshenv": true,
57
+ ".zlogin": true,
58
+ ".zprofile": true,
59
+ ".zlogout": true,
60
+ ".python_history": true,
61
+ ".lightningignore": true,
62
+ ".nv": true,
63
+ ".docker": true,
64
+ ".jupyter": true,
65
+ ".lightning": true,
66
+ ".vscode": true,
67
+ ".vscode/**": true,
68
+ ".wget-hsts": true,
69
+ ".vscode-server/**": true,
70
+ ".vscode-server-insiders/**": true,
71
+ ".cursor-server/**": true,
72
+ ".windsurf-server/**": true,
73
+ ".tmplaigit": true,
74
+ ".viminfo": true,
75
+ ".idea/**": true,
76
+ ".java/**": true,
77
+ ".dotnet/**": true,
78
+ "**/node_modules/**": true,
79
+ "**/__pycache__/**": true,
80
+ "**/.tmp_versions/**": true,
81
+ "**/.yarn/**": true,
82
+ "**/__pypackages__/**": true,
83
+ "**/*_cache": true,
84
+ "**/.lightning-job-artifacts": true
85
+ },
86
+ "files.associations": {
87
+ "*.studiorc": "shellscript"
88
+ },
89
+ "files.enableTrash": false
90
+ }
bart/data/mni_train_tokenise.txt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a1136eec92536c02400a10defedb22e399c1232c5efffc3393c93450b881c622
3
+ size 615472997
data-bin/dict.eng_Latn.txt ADDED
The diff for this file is too large to render. See raw diff
 
data-bin/dict.mni_Beng.txt ADDED
The diff for this file is too large to render. See raw diff
 
data-bin/preprocess.log ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Namespace(aim_repo=None, aim_run_hash=None, align_suffix=None, alignfile=None, all_gather_list_size=16384, amp=False, amp_batch_retries=2, amp_init_scale=128, amp_scale_window=None, azureml_logging=False, bf16=False, bpe=None, cpu=False, criterion='cross_entropy', dataset_impl='mmap', destdir='data-bin', dict_only=False, empty_cache_freq=0, fp16=False, fp16_init_scale=128, fp16_no_flatten_grads=False, fp16_scale_tolerance=0.0, fp16_scale_window=None, joined_dictionary=False, log_file=None, log_format=None, log_interval=100, lr_scheduler='fixed', memory_efficient_bf16=False, memory_efficient_fp16=False, min_loss_scale=0.0001, model_parallel_size=1, no_progress_bar=False, nwordssrc=-1, nwordstgt=-1, on_cpu_convert_precision=False, only_source=False, optimizer=None, padding_factor=8, plasma_path='/tmp/plasma', profile=False, quantization_config_path=None, reset_logging=False, scoring='bleu', seed=1, source_lang='eng_Latn', srcdict=None, suppress_crashes=False, target_lang='mni_Beng', task='translation', tensorboard_logdir=None, testpref='data/test', tgtdict=None, threshold_loss_scale=None, thresholdsrc=0, thresholdtgt=0, tokenizer=None, tpu=False, trainpref='data/train', use_plasma_view=False, user_dir=None, validpref='data/dev', wandb_project=None, workers=1)
2
+ [eng_Latn] Dictionary: 104488 types
3
+ [eng_Latn] data/train.eng_Latn: 139266 sents, 2851668 tokens, 0.0% replaced (by <unk>)
4
+ [eng_Latn] Dictionary: 104488 types
5
+ [eng_Latn] data/dev.eng_Latn: 2404 sents, 48999 tokens, 1.72% replaced (by <unk>)
6
+ [eng_Latn] Dictionary: 104488 types
7
+ [eng_Latn] data/test.eng_Latn: 1183 sents, 23873 tokens, 1.63% replaced (by <unk>)
8
+ [mni_Beng] Dictionary: 185504 types
9
+ [mni_Beng] data/train.mni_Beng: 139266 sents, 2418773 tokens, 0.0% replaced (by <unk>)
10
+ [mni_Beng] Dictionary: 185504 types
11
+ [mni_Beng] data/dev.mni_Beng: 2404 sents, 41351 tokens, 3.87% replaced (by <unk>)
12
+ [mni_Beng] Dictionary: 185504 types
13
+ [mni_Beng] data/test.mni_Beng: 1183 sents, 20030 tokens, 4.03% replaced (by <unk>)
14
+ Wrote preprocessed data to data-bin
data-bin/test.eng_Latn-mni_Beng.eng_Latn.idx ADDED
Binary file (14.2 kB). View file
 
data-bin/test.eng_Latn-mni_Beng.mni_Beng.idx ADDED
Binary file (14.2 kB). View file
 
data-bin/valid.eng_Latn-mni_Beng.eng_Latn.idx ADDED
Binary file (28.9 kB). View file
 
data-bin/valid.eng_Latn-mni_Beng.mni_Beng.idx ADDED
Binary file (28.9 kB). View file
 
data/checkpoints/checkpoint20.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e3361510c3efb318842eb60fc657c10d6d6ba8317acb38fc212f8dacd52efd20
3
+ size 2374486540
data/fairseq/.github/CODEOWNERS ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Setting up CODEOWNERS for UST related codebase
2
+ # Documentation for open sourced models relevant to UST
3
+ examples/speech_to_text @kahne @sravyapopuri388 @jmp84
4
+ examples/speech_to_speech @an918tw @sravyapopuri388 @jmp84
5
+ examples/speech_synthesis @kahne @jmp84
6
+ examples/simultaneous_translation @kahne @jmp84
7
+ examples/speech_text_joint_to_text @yuntang @jmp84
8
+
9
+ # Speech related models relevant to UST
10
+ fairseq/models/speech_to_speech @sravyapopuri388 @jmp84
11
+ fairseq/models/speech_to_text @kahne @sravyapopuri388 @jmp84
12
+ fairseq/models/text_to_speech @kahne @jmp84
13
+
14
+ # CONFORMER IMPLEMENTATION
15
+ fairseq/modules/conformer_layer.py @sravyapopuri388 @jmp84
16
+ fairseq/modules/espnet_multihead_attention.py @sravyapopuri388 @jmp84
17
+ fairseq/modules/rotary_positional_embedding.py @sravyapopuri388 @jmp84
18
+ fairseq/modules/positional_encoding.py @sravyapopuri388 @jmp84
19
+
20
+ # Machine Translation/NLLB
21
+ fairseq/tasks/translation.py @gwenzek
data/fairseq/.github/ISSUE_TEMPLATE.md ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ ## 👉 [Please follow one of these issue templates](https://github.com/pytorch/fairseq/issues/new/choose) 👈
2
+
3
+ Note: to keep the backlog clean and actionable, issues may be immediately closed if they do not follow one of the above issue templates.
data/fairseq/.github/ISSUE_TEMPLATE/bug_report.md ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ name: 🐛 Bug Report
3
+ about: Submit a bug report to help us improve
4
+ labels: 'bug, needs triage'
5
+ ---
6
+
7
+ ## 🐛 Bug
8
+
9
+ <!-- A clear and concise description of what the bug is. -->
10
+
11
+ ### To Reproduce
12
+
13
+ Steps to reproduce the behavior (**always include the command you ran**):
14
+
15
+ 1. Run cmd '....'
16
+ 2. See error
17
+
18
+ <!-- If you have a code sample, error messages, stack traces, please provide it here as well -->
19
+
20
+
21
+ #### Code sample
22
+ <!-- Ideally attach a minimal code sample to reproduce the decried issue.
23
+ Minimal means having the shortest code but still preserving the bug. -->
24
+
25
+ ### Expected behavior
26
+
27
+ <!-- A clear and concise description of what you expected to happen. -->
28
+
29
+ ### Environment
30
+
31
+ - fairseq Version (e.g., 1.0 or main):
32
+ - PyTorch Version (e.g., 1.0)
33
+ - OS (e.g., Linux):
34
+ - How you installed fairseq (`pip`, source):
35
+ - Build command you used (if compiling from source):
36
+ - Python version:
37
+ - CUDA/cuDNN version:
38
+ - GPU models and configuration:
39
+ - Any other relevant information:
40
+
41
+ ### Additional context
42
+
43
+ <!-- Add any other context about the problem here. -->
data/fairseq/.github/ISSUE_TEMPLATE/documentation.md ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ name: 📚 Documentation/Typos
3
+ about: Report an issue related to documentation or a typo
4
+ labels: 'documentation, needs triage'
5
+ ---
6
+
7
+ ## 📚 Documentation
8
+
9
+ For typos and doc fixes, please go ahead and:
10
+
11
+ 1. Create an issue.
12
+ 2. Fix the typo.
13
+ 3. Submit a PR.
14
+
15
+ Thanks!
data/fairseq/.github/ISSUE_TEMPLATE/feature_request.md ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ name: 🚀 Feature Request
3
+ about: Submit a proposal/request for a new feature
4
+ labels: 'enhancement, help wanted, needs triage'
5
+ ---
6
+
7
+ ## 🚀 Feature Request
8
+ <!-- A clear and concise description of the feature proposal -->
9
+
10
+ ### Motivation
11
+
12
+ <!-- Please outline the motivation for the proposal. Is your feature request related to a problem? e.g., I'm always frustrated when [...]. If this is related to another GitHub issue, please link here too -->
13
+
14
+ ### Pitch
15
+
16
+ <!-- A clear and concise description of what you want to happen. -->
17
+
18
+ ### Alternatives
19
+
20
+ <!-- A clear and concise description of any alternative solutions or features you've considered, if any. -->
21
+
22
+ ### Additional context
23
+
24
+ <!-- Add any other context or screenshots about the feature request here. -->
data/fairseq/.github/ISSUE_TEMPLATE/how-to-question.md ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ name: ❓ Questions/Help
3
+ about: If you have questions, please first search existing issues and docs
4
+ labels: 'question, needs triage'
5
+ ---
6
+
7
+ ## ❓ Questions and Help
8
+
9
+ ### Before asking:
10
+ 1. search the issues.
11
+ 2. search the docs.
12
+
13
+ <!-- If you still can't find what you need: -->
14
+
15
+ #### What is your question?
16
+
17
+ #### Code
18
+
19
+ <!-- Please paste a code snippet if your question requires it! -->
20
+
21
+ #### What have you tried?
22
+
23
+ #### What's your environment?
24
+
25
+ - fairseq Version (e.g., 1.0 or main):
26
+ - PyTorch Version (e.g., 1.0)
27
+ - OS (e.g., Linux):
28
+ - How you installed fairseq (`pip`, source):
29
+ - Build command you used (if compiling from source):
30
+ - Python version:
31
+ - CUDA/cuDNN version:
32
+ - GPU models and configuration:
33
+ - Any other relevant information:
data/fairseq/.github/PULL_REQUEST_TEMPLATE.md ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Before submitting
2
+
3
+ - [ ] Was this discussed/approved via a Github issue? (no need for typos, doc improvements)
4
+ - [ ] Did you read the [contributor guideline](https://github.com/pytorch/fairseq/blob/main/CONTRIBUTING.md)?
5
+ - [ ] Did you make sure to update the docs?
6
+ - [ ] Did you write any new necessary tests?
7
+
8
+ ## What does this PR do?
9
+ Fixes # (issue).
10
+
11
+ ## PR review
12
+ Anyone in the community is free to review the PR once the tests have passed.
13
+ If we didn't discuss your PR in Github issues there's a high chance it will not be merged.
14
+
15
+ ## Did you have fun?
16
+ Make sure you had fun coding 🙃
data/fairseq/.github/stale.yml ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Configuration for probot-stale - https://github.com/probot/stale
2
+ # Mostly copied from github.com/facebook/react/blob/master/.github/stale.yml
3
+ # Number of days of inactivity before an issue becomes stale
4
+ daysUntilStale: 90
5
+ # Number of days of inactivity before a stale issue is closed
6
+ daysUntilClose: 7
7
+ # Issues with these labels will never be considered stale
8
+ exemptLabels:
9
+ - bug
10
+ # Label to use when marking an issue as stale
11
+ staleLabel: stale
12
+ issues:
13
+ # Comment to post when marking an issue as stale.
14
+ markComment: >
15
+ This issue has been automatically marked as stale.
16
+ **If this issue is still affecting you, please leave any comment** (for example, "bump"), and we'll keep it open.
17
+ We are sorry that we haven't been able to prioritize it yet. If you have any new additional information, please include it with your comment!
18
+ # Comment to post when closing a stale issue.
19
+ closeComment: >
20
+ Closing this issue after a prolonged period of inactivity. If this issue is still present in the latest release, please create a new issue with up-to-date information. Thank you!
21
+ pulls:
22
+ # Comment to post when marking a pull request as stale.
23
+ markComment: >
24
+ This pull request has been automatically marked as stale.
25
+ **If this pull request is still relevant, please leave any comment** (for example, "bump"), and we'll keep it open.
26
+ We are sorry that we haven't been able to prioritize reviewing it yet. Your contribution is very much appreciated.
27
+ # Comment to post when closing a stale pull request.
28
+ closeComment: >
29
+ Closing this pull request after a prolonged period of inactivity. If this issue is still present in the latest release, please ask for this pull request to be reopened. Thank you!
30
+
data/fairseq/.github/workflows/build.yml ADDED
@@ -0,0 +1,81 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ name: build
2
+
3
+ on:
4
+ # Trigger the workflow on push to main or any pull request
5
+ push:
6
+ branches:
7
+ - main
8
+ pull_request:
9
+
10
+ jobs:
11
+ build:
12
+
13
+ strategy:
14
+ max-parallel: 4
15
+ matrix:
16
+ platform: [ubuntu-latest, macos-latest]
17
+ python-version: [3.8, 3.9]
18
+
19
+ runs-on: ${{ matrix.platform }}
20
+
21
+ steps:
22
+ - uses: actions/checkout@v2
23
+
24
+ - name: Set up Python ${{ matrix.python-version }}
25
+ uses: actions/setup-python@v2
26
+ with:
27
+ python-version: ${{ matrix.python-version }}
28
+
29
+ - name: Conditionally install pytorch
30
+ if: matrix.platform == 'windows-latest'
31
+ run: pip3 install torch -f https://download.pytorch.org/whl/torch_stable.html
32
+
33
+ - name: Install locally
34
+ run: |
35
+ python -m pip install --upgrade pip
36
+ git submodule update --init --recursive
37
+ python -m pip install .
38
+
39
+ - name: Check installation
40
+ working-directory: /tmp
41
+ run: python $GITHUB_WORKSPACE/scripts/check_installation.py
42
+
43
+ - name: Install optional test requirements
44
+ run: |
45
+ python -m pip install '.[dev,docs]'
46
+ python -m pip install iopath transformers pyarrow
47
+ python -m pip install git+https://github.com/facebookresearch/fairscale.git@main
48
+ python -m pip install pygit2 pgzip
49
+
50
+ - name: Install xformers for Macos
51
+ if: matrix.platform == 'macos-latest'
52
+ run: |
53
+ brew install llvm libomp
54
+ CC=/usr/local/opt/llvm/bin/clang CXX=clang++ pip install git+https://github.com/facebookresearch/xformers.git@main
55
+
56
+ - name: Install xformers for non-MacOS
57
+ if: matrix.platform != 'macos-latest'
58
+ run: |
59
+ python -m pip install --progress-bar off git+https://github.com/facebookresearch/xformers.git@main
60
+
61
+ - name: Lint with black
62
+ run: black --check --diff .
63
+
64
+ - name: Lint with flake8
65
+ run: |
66
+ # stop the build if there are Python syntax errors or undefined names
67
+ flake8 . --count --select=E9,F63,F7,F82 --show-source --statistics
68
+ # exit-zero treats all errors as warnings. The GitHub editor is 127 chars wide
69
+ flake8 . --count --exit-zero --max-complexity=10 --max-line-length=127 --statistics
70
+
71
+ - name: Build doc
72
+ run: make singlehtml
73
+ working-directory: docs/
74
+
75
+ - name: Run tests
76
+ # When installing in non-editable mode, the .so files will be generated in 'site-packages/fairseq'.
77
+ # But by default, pytest import machinery will load local fairseq, and won't see the .so.
78
+ # Use --import-mode=append to favorize the 'site-packages/fairseq'.
79
+ # https://docs.pytest.org/en/7.1.x/explanation/pythonpath.html
80
+ run: pytest --import-mode=append -vvv tests/
81
+
data/fairseq/.github/workflows/depreview.yml ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ name: 'Dependency Review'
2
+ on: [pull_request]
3
+
4
+ permissions:
5
+ contents: read
6
+
7
+ jobs:
8
+ dependency-review:
9
+ runs-on: ubuntu-latest
10
+ steps:
11
+ - name: 'Checkout Repository'
12
+ uses: actions/checkout@v4
13
+ - name: Dependency Review
14
+ uses: actions/dependency-review-action@v4
data/fairseq/.github/workflows/release.yml ADDED
@@ -0,0 +1,161 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ name: Fairseq Release
2
+
3
+ on:
4
+ workflow_dispatch:
5
+ inputs:
6
+ name:
7
+ description: 'Release Type'
8
+ default: 'patch'
9
+ required: true
10
+
11
+ jobs:
12
+
13
+ get_next_version:
14
+ runs-on: ubuntu-latest
15
+ steps:
16
+ - name: checkout-repo-content
17
+ uses: actions/checkout@v2
18
+
19
+ - name: setup-python
20
+ uses: actions/setup-python@v2
21
+ with:
22
+ python-version: 3.8
23
+
24
+ - name: get next version and tag
25
+ id: get-next-version-and-tag
26
+ run: |
27
+ output=$(python3 release_utils.py --release-type ${{ github.event.inputs.name }})
28
+ echo $output
29
+ new_version=$(echo $output | awk '{print $1}')
30
+ new_tag=$(echo $output | awk '{print $2}')
31
+ echo "new version is $new_version"
32
+ echo "new tag is $new_tag"
33
+ echo ::set-output name=version::$new_version
34
+ echo ::set-output name=tag::$new_tag
35
+ echo ::set-output name=branch_name::$new_version-release
36
+ echo "NEW_TAG=$new_tag" >> $GITHUB_ENV
37
+ echo "NEW_BRANCH=$new_version-release" >> $GITHUB_ENV
38
+
39
+
40
+ # update the version number in version.txt
41
+ - name: update version
42
+ id: update-version
43
+ run : |
44
+ echo "current folder = $PWD"
45
+ echo "current branch = $(git branch --show-current)"
46
+ output=$(python3 release_utils.py --release-type ${{ github.event.inputs.name }} --update-version)
47
+
48
+ - name: add and commit
49
+ uses: EndBug/add-and-commit@v9
50
+ with:
51
+ author_name: ${{ secrets.AUTHOR_NAME }}
52
+ author_email: ${{ secrets.AUTHOR_EMAIL }}
53
+
54
+ # TODO: change this to main once shipit is disabled.
55
+ new_branch: '${{ env.NEW_BRANCH }}'
56
+ default_author: github_actor
57
+ message: '${{ env.NEW_TAG }} release'
58
+ pathspec_error_handling: exitAtEnd
59
+
60
+ # Arguments for the git pull command. Use NO-PULL to avoid the action pulling at all.
61
+ # pull: 'NO-PULL'
62
+ tag: '${{ env.NEW_TAG }}'
63
+
64
+ outputs:
65
+ new_version: ${{ steps.get-next-version-and-tag.outputs.version }}
66
+ new_tag: ${{ steps.get-next-version-and-tag.outputs.tag }}
67
+ branch_name: ${{ steps.get-next-version-and-tag.outputs.branch_name }}
68
+
69
+ create_sdist:
70
+ runs-on: ubuntu-latest
71
+ name: Create Source Distribution
72
+ needs: get_next_version
73
+ steps:
74
+ - uses: actions/checkout@v3
75
+ with:
76
+ ref: ${{ needs.get_next_version.outputs.branch_name }}
77
+
78
+ - name: Install Python
79
+ uses: actions/setup-python@v2
80
+ with:
81
+ python-version: '3.8'
82
+
83
+ - name: Upgrade pip
84
+ run: |
85
+ python3 -m pip install --upgrade pip
86
+
87
+ - name: Create Source Distribution
88
+ run: |
89
+ python3 -m pip install setuptools wheel twine torch
90
+ python3 setup.py sdist
91
+
92
+ - uses: actions/upload-artifact@v2
93
+ with:
94
+ path: dist/*.tar.gz
95
+
96
+ build_wheels:
97
+ name: Build wheels on ${{ matrix.os }}
98
+ runs-on: ${{ matrix.os }}
99
+ needs: get_next_version
100
+ strategy:
101
+ matrix:
102
+ os: [ubuntu-latest, macos-latest]
103
+
104
+ steps:
105
+ - uses: actions/checkout@v3
106
+ with:
107
+ ref: ${{ needs.get_next_version.outputs.branch_name }}
108
+
109
+ - name: Install Python
110
+ uses: actions/setup-python@v2
111
+ with:
112
+ python-version: '3.8'
113
+
114
+ - name: Upgrade pip
115
+ run: |
116
+ python3 -m pip install --upgrade pip
117
+
118
+ - name: Install cibuildwheel
119
+ run: |
120
+ python3 -m pip install cibuildwheel
121
+
122
+ - name: Build wheels for CPython
123
+ run: |
124
+ python3 -m cibuildwheel --output-dir dist
125
+ env:
126
+ CIBW_BUILD: "cp38-*64"
127
+ CIBW_MANYLINUX_X86_64_IMAGE: manylinux1
128
+ CIBW_BEFORE_BUILD: git submodule update --init --recursive && pip install .
129
+ # Install system library
130
+ CIBW_BEFORE_BUILD_LINUX: (yum install -y libffi-devel || apt-get install -y libffi-devel || apk add --update --no-cache libffi-devel || true) && (yum install -y libc6 || apt-get install -y libc6 || apk add --update --no-cache libc6 || true)
131
+ CIBW_ENVIRONMENT: "PIP_ONLY_BINARY=numpy"
132
+ CIBW_SKIP: "*musllinux*"
133
+
134
+ - uses: actions/upload-artifact@v2
135
+ with:
136
+ path: dist
137
+
138
+ upload:
139
+ name: Upload to PyPi and create release
140
+ runs-on: ubuntu-latest
141
+ needs: [build_wheels, create_sdist, get_next_version]
142
+ steps:
143
+ - uses: actions/download-artifact@v2
144
+ with:
145
+ name: artifact
146
+ path: dist
147
+
148
+ # build the PyPI package and upload it
149
+ - name: upload
150
+ env:
151
+ TWINE_USERNAME: ${{ secrets.PYPI_USERNAME }}
152
+ TWINE_PASSWORD: ${{ secrets.PYPI_PASSWORD }}
153
+ run: |
154
+ pip install setuptools wheel twine
155
+ python3 -m twine upload --repository pypi dist/*
156
+
157
+ # create the release on github
158
+ - name: create release on github
159
+ uses: ncipollo/release-action@v1
160
+ with:
161
+ tag: '${{ needs.get_next_version.outputs.new_tag }}'
data/fairseq/fairseq/clib/cuda/ngram_repeat_block_cuda_kernel.cu ADDED
@@ -0,0 +1,82 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ /*
2
+ Copyright (c) Microsoft Corporation.
3
+ Licensed under the MIT License.
4
+ */
5
+
6
+ /*
7
+ Kernel implementation for blocking repeated n-grams.
8
+ */
9
+
10
+ #include <cuda.h>
11
+ #include <cuda_runtime.h>
12
+ #include <math.h>
13
+ #include <torch/extension.h>
14
+ #include <vector>
15
+
16
+ // Ban repeated ngrams of length = 'no_repeat_ngram_size'
17
+ __global__ void banRepeatedTokens(
18
+ long* __restrict__ tokens,
19
+ float* __restrict__ lprobs,
20
+ int max_predict_len,
21
+ int vocab_size,
22
+ int no_repeat_ngram_size) {
23
+ auto row = blockIdx.x;
24
+ auto col = threadIdx.x;
25
+ auto start = row * (max_predict_len) + col;
26
+ // Each thread compares ngram starting from
27
+ // thread index with final ngram starting from
28
+ // step - no_repeat_ngram_size +2
29
+ auto check_start_pos = blockDim.x;
30
+ auto lprob_start = row * vocab_size;
31
+ bool is_banned = true;
32
+ extern __shared__ long tokens_shm[];
33
+ tokens_shm[col] = tokens[start];
34
+ if (col == blockDim.x - 1) {
35
+ for (int i = 1; i < no_repeat_ngram_size; i++) {
36
+ if (col + i < max_predict_len) {
37
+ tokens_shm[col + i] = tokens[start + i];
38
+ }
39
+ }
40
+ }
41
+ __syncthreads();
42
+
43
+ for (int k = 0; k < no_repeat_ngram_size - 1; k++) {
44
+ if (tokens_shm[col + k] != tokens_shm[check_start_pos + k]) {
45
+ is_banned = false;
46
+ }
47
+ }
48
+ if (is_banned == true) {
49
+ auto token_to_be_banned = tokens_shm[col + no_repeat_ngram_size - 1];
50
+ lprobs[lprob_start + token_to_be_banned] = -INFINITY;
51
+ }
52
+ }
53
+
54
+ // Allocate blocks and threads based on
55
+ // batch size and sequence length and launch
56
+ // kernel
57
+ torch::Tensor ngram_repeat_block_cuda_forward(
58
+ const torch::Tensor tokens,
59
+ torch::Tensor lprobs,
60
+ int bsz,
61
+ int step,
62
+ int beam_size,
63
+ int no_repeat_ngram_size) {
64
+ int threads = step - no_repeat_ngram_size + 2;
65
+ if (threads <= 0)
66
+ return lprobs;
67
+ int max_predict_len = tokens.size(1);
68
+ int vocab_size = lprobs.size(1);
69
+ auto token_ptr = tokens.data_ptr<long>();
70
+ auto lprob_ptr = lprobs.data_ptr<float>();
71
+ int blocks = bsz * beam_size;
72
+ int shared_mem_size = (step + 1) * sizeof(long);
73
+
74
+ // Launching N blocks where N is number of samples in a batch (beams*bsz)
75
+ // Launching T threads where T is number of previous ngrams in a sample
76
+ // Allocating shared mem per block for fastser access of input tokens since
77
+ // each token will be accessed N times to compare with current Ngram where
78
+ // N is Ngram size.
79
+ banRepeatedTokens<<<blocks, threads, shared_mem_size>>>(
80
+ token_ptr, lprob_ptr, max_predict_len, vocab_size, no_repeat_ngram_size);
81
+ return lprobs;
82
+ }
data/fairseq/fairseq/clib/libbase/balanced_assignment.cpp ADDED
@@ -0,0 +1,109 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ /**
2
+ * Copyright 2017-present, Facebook, Inc.
3
+ * All rights reserved.
4
+ *
5
+ * This source code is licensed under the license found in the
6
+ * LICENSE file in the root directory of this source tree.
7
+ */
8
+
9
+ /*
10
+ C++ code for solving the linear assignment problem.
11
+ Based on the Auction Algorithm from
12
+ https://dspace.mit.edu/bitstream/handle/1721.1/3265/P-2108-26912652.pdf and the
13
+ implementation from: https://github.com/bkj/auction-lap Adapted to be more
14
+ efficient when each worker is looking for k jobs instead of 1.
15
+ */
16
+ #include <torch/extension.h>
17
+ #include <iostream>
18
+ using namespace torch::indexing;
19
+ torch::Tensor balanced_assignment(torch::Tensor job_and_worker_to_score) {
20
+ int max_iterations = 100;
21
+ torch::Tensor epsilon =
22
+ (job_and_worker_to_score.max() - job_and_worker_to_score.min()) / 50;
23
+ epsilon.clamp_min_(1e-04);
24
+ torch::Tensor worker_and_job_to_score =
25
+ job_and_worker_to_score.detach().transpose(0, 1).contiguous();
26
+ int num_workers = worker_and_job_to_score.size(0);
27
+ int num_jobs = worker_and_job_to_score.size(1);
28
+ auto device = worker_and_job_to_score.device();
29
+ int jobs_per_worker = num_jobs / num_workers;
30
+ torch::Tensor value = worker_and_job_to_score.clone();
31
+ int counter = 0;
32
+ torch::Tensor max_value = worker_and_job_to_score.max();
33
+
34
+ torch::Tensor bid_indices;
35
+ torch::Tensor cost = worker_and_job_to_score.new_zeros({1, num_jobs});
36
+ torch::Tensor bids =
37
+ worker_and_job_to_score.new_empty({num_workers, num_jobs});
38
+ torch::Tensor bid_increments =
39
+ worker_and_job_to_score.new_empty({num_workers, jobs_per_worker});
40
+ torch::Tensor top_values =
41
+ worker_and_job_to_score.new_empty({num_workers, jobs_per_worker + 1});
42
+ torch::Tensor high_bids = worker_and_job_to_score.new_empty({num_jobs});
43
+
44
+ torch::Tensor top_index = top_values.to(torch::kLong);
45
+ torch::Tensor high_bidders = top_index.new_empty({num_jobs});
46
+ torch::Tensor have_bids = high_bidders.to(torch::kBool);
47
+ torch::Tensor jobs_indices =
48
+ torch::arange({num_jobs}, torch::dtype(torch::kLong).device(device));
49
+ torch::Tensor true_tensor =
50
+ torch::ones({1}, torch::dtype(torch::kBool).device(device));
51
+
52
+ while (true) {
53
+ bids.zero_();
54
+ torch::topk_out(top_values, top_index, value, jobs_per_worker + 1, 1);
55
+
56
+ // Each worker bids the difference in value between that job and the k+1th
57
+ // job
58
+ torch::sub_out(
59
+ bid_increments,
60
+ top_values.index({Slice(None, None), Slice(0, jobs_per_worker)}),
61
+ top_values.index({Slice(None, None), jobs_per_worker}).unsqueeze(1));
62
+
63
+ bid_increments.add_(epsilon);
64
+ bids.scatter_(
65
+ 1,
66
+ top_index.index({Slice(None, None), Slice(0, jobs_per_worker)}),
67
+ bid_increments);
68
+
69
+ if (counter < max_iterations && counter > 0) {
70
+ // Put in a minimal bid to retain items from the last round if no-one else
71
+ // bids for them this round
72
+ bids.view(-1).index_put_({bid_indices}, epsilon);
73
+ }
74
+
75
+ // Find the highest bidding worker per job
76
+ torch::max_out(high_bids, high_bidders, bids, 0);
77
+ torch::gt_out(have_bids, high_bids, 0);
78
+
79
+ if (have_bids.all().item<bool>()) {
80
+ // All jobs were bid for
81
+ break;
82
+ }
83
+
84
+ // Make popular items more expensive
85
+ cost.add_(high_bids);
86
+ torch::sub_out(value, worker_and_job_to_score, cost);
87
+
88
+ bid_indices = ((high_bidders * num_jobs) + jobs_indices).index({have_bids});
89
+
90
+ if (counter < max_iterations) {
91
+ // Make sure that this item will be in the winning worker's top-k next
92
+ // time.
93
+ value.view(-1).index_put_({bid_indices}, max_value);
94
+ } else {
95
+ // Suboptimal approximation that converges quickly from current solution
96
+ value.view(-1).index_put_(
97
+ {bid_indices}, worker_and_job_to_score.view(-1).index({bid_indices}));
98
+ }
99
+
100
+ counter += 1;
101
+ }
102
+
103
+ return top_index.index({Slice(None, None), Slice(0, jobs_per_worker)})
104
+ .reshape(-1);
105
+ }
106
+
107
+ PYBIND11_MODULE(TORCH_EXTENSION_NAME, m) {
108
+ m.def("balanced_assignment", &balanced_assignment, "Balanced Assignment");
109
+ }
data/fairseq/fairseq/clib/libnat/edit_dist.cpp ADDED
@@ -0,0 +1,231 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ /**
2
+ * Copyright 2017-present, Facebook, Inc.
3
+ * All rights reserved.
4
+ *
5
+ * This source code is licensed under the license found in the
6
+ * LICENSE file in the root directory of this source tree.
7
+ */
8
+
9
+ #include <pybind11/detail/common.h>
10
+ #include <pybind11/pybind11.h>
11
+ #include <torch/torch.h> // @manual=//caffe2:torch_extension
12
+ #include <algorithm>
13
+ #include <cstdint>
14
+ #include <iosfwd>
15
+ #include <memory>
16
+ #include <new>
17
+ #include <string>
18
+ #include <utility>
19
+ #include <vector>
20
+
21
+ using namespace ::std;
22
+
23
+ vector<vector<uint32_t>> edit_distance2_with_dp(
24
+ vector<uint32_t>& x,
25
+ vector<uint32_t>& y) {
26
+ uint32_t lx = x.size();
27
+ uint32_t ly = y.size();
28
+ vector<vector<uint32_t>> d(lx + 1, vector<uint32_t>(ly + 1));
29
+ for (uint32_t i = 0; i < lx + 1; i++) {
30
+ d[i][0] = i;
31
+ }
32
+ for (uint32_t j = 0; j < ly + 1; j++) {
33
+ d[0][j] = j;
34
+ }
35
+ for (uint32_t i = 1; i < lx + 1; i++) {
36
+ for (uint32_t j = 1; j < ly + 1; j++) {
37
+ d[i][j] =
38
+ min(min(d[i - 1][j], d[i][j - 1]) + 1,
39
+ d[i - 1][j - 1] + 2 * (x.at(i - 1) == y.at(j - 1) ? 0 : 1));
40
+ }
41
+ }
42
+ return d;
43
+ }
44
+
45
+ vector<vector<uint32_t>> edit_distance2_backtracking(
46
+ vector<vector<uint32_t>>& d,
47
+ vector<uint32_t>& x,
48
+ vector<uint32_t>& y,
49
+ uint32_t terminal_symbol) {
50
+ vector<uint32_t> seq;
51
+ vector<vector<uint32_t>> edit_seqs(x.size() + 2, vector<uint32_t>());
52
+ /*
53
+ edit_seqs:
54
+ 0~x.size() cell is the insertion sequences
55
+ last cell is the delete sequence
56
+ */
57
+
58
+ if (x.size() == 0) {
59
+ edit_seqs.at(0) = y;
60
+ return edit_seqs;
61
+ }
62
+
63
+ uint32_t i = d.size() - 1;
64
+ uint32_t j = d.at(0).size() - 1;
65
+
66
+ while ((i >= 0) && (j >= 0)) {
67
+ if ((i == 0) && (j == 0)) {
68
+ break;
69
+ }
70
+
71
+ if ((j > 0) && (d.at(i).at(j - 1) < d.at(i).at(j))) {
72
+ seq.push_back(1); // insert
73
+ seq.push_back(y.at(j - 1));
74
+ j--;
75
+ } else if ((i > 0) && (d.at(i - 1).at(j) < d.at(i).at(j))) {
76
+ seq.push_back(2); // delete
77
+ seq.push_back(x.at(i - 1));
78
+ i--;
79
+ } else {
80
+ seq.push_back(3); // keep
81
+ seq.push_back(x.at(i - 1));
82
+ i--;
83
+ j--;
84
+ }
85
+ }
86
+
87
+ uint32_t prev_op, op, s, word;
88
+ prev_op = 0, s = 0;
89
+ for (uint32_t k = 0; k < seq.size() / 2; k++) {
90
+ op = seq.at(seq.size() - 2 * k - 2);
91
+ word = seq.at(seq.size() - 2 * k - 1);
92
+ if (prev_op != 1) {
93
+ s++;
94
+ }
95
+ if (op == 1) // insert
96
+ {
97
+ edit_seqs.at(s - 1).push_back(word);
98
+ } else if (op == 2) // delete
99
+ {
100
+ edit_seqs.at(x.size() + 1).push_back(1);
101
+ } else {
102
+ edit_seqs.at(x.size() + 1).push_back(0);
103
+ }
104
+
105
+ prev_op = op;
106
+ }
107
+
108
+ for (uint32_t k = 0; k < edit_seqs.size(); k++) {
109
+ if (edit_seqs[k].size() == 0) {
110
+ edit_seqs[k].push_back(terminal_symbol);
111
+ }
112
+ }
113
+ return edit_seqs;
114
+ }
115
+
116
+ vector<vector<uint32_t>> edit_distance2_backtracking_with_delete(
117
+ vector<vector<uint32_t>>& d,
118
+ vector<uint32_t>& x,
119
+ vector<uint32_t>& y,
120
+ uint32_t terminal_symbol,
121
+ uint32_t deletion_symbol) {
122
+ vector<uint32_t> seq;
123
+ vector<vector<uint32_t>> edit_seqs(x.size() + 1, vector<uint32_t>());
124
+ /*
125
+ edit_seqs:
126
+ 0~x.size() cell is the insertion sequences
127
+ last cell is the delete sequence
128
+ */
129
+
130
+ if (x.size() == 0) {
131
+ edit_seqs.at(0) = y;
132
+ return edit_seqs;
133
+ }
134
+
135
+ uint32_t i = d.size() - 1;
136
+ uint32_t j = d.at(0).size() - 1;
137
+
138
+ while ((i >= 0) && (j >= 0)) {
139
+ if ((i == 0) && (j == 0)) {
140
+ break;
141
+ }
142
+
143
+ if ((j > 0) && (d.at(i).at(j - 1) < d.at(i).at(j))) {
144
+ seq.push_back(1); // insert
145
+ seq.push_back(y.at(j - 1));
146
+ j--;
147
+ } else if ((i > 0) && (d.at(i - 1).at(j) < d.at(i).at(j))) {
148
+ seq.push_back(2); // delete
149
+ seq.push_back(x.at(i - 1));
150
+ i--;
151
+ } else {
152
+ seq.push_back(3); // keep
153
+ seq.push_back(x.at(i - 1));
154
+ i--;
155
+ j--;
156
+ }
157
+ }
158
+
159
+ uint32_t prev_op, op, s, word;
160
+ prev_op = 0, s = 0;
161
+ for (uint32_t k = 0; k < seq.size() / 2; k++) {
162
+ op = seq.at(seq.size() - 2 * k - 2);
163
+ word = seq.at(seq.size() - 2 * k - 1);
164
+ if (prev_op != 1) {
165
+ s++;
166
+ }
167
+ if (op == 1) // insert
168
+ {
169
+ edit_seqs.at(s - 1).push_back(word);
170
+ } else if (op == 2) // delete
171
+ {
172
+ edit_seqs.at(s - 1).push_back(deletion_symbol);
173
+ }
174
+
175
+ prev_op = op;
176
+ }
177
+
178
+ for (uint32_t k = 0; k < edit_seqs.size(); k++) {
179
+ if (edit_seqs.at(k).size() == 0) {
180
+ edit_seqs.at(k).push_back(terminal_symbol);
181
+ }
182
+ }
183
+ return edit_seqs;
184
+ }
185
+
186
+ vector<uint32_t> compute_ed2(
187
+ vector<vector<uint32_t>>& xs,
188
+ vector<vector<uint32_t>>& ys) {
189
+ vector<uint32_t> distances(xs.size());
190
+ for (uint32_t i = 0; i < xs.size(); i++) {
191
+ vector<vector<uint32_t>> d = edit_distance2_with_dp(xs.at(i), ys.at(i));
192
+ distances.at(i) = d.at(xs.at(i).size()).at(ys.at(i).size());
193
+ }
194
+ return distances;
195
+ }
196
+
197
+ vector<vector<vector<uint32_t>>> suggested_ed2_path(
198
+ vector<vector<uint32_t>>& xs,
199
+ vector<vector<uint32_t>>& ys,
200
+ uint32_t terminal_symbol) {
201
+ vector<vector<vector<uint32_t>>> seq(xs.size());
202
+ for (uint32_t i = 0; i < xs.size(); i++) {
203
+ vector<vector<uint32_t>> d = edit_distance2_with_dp(xs.at(i), ys.at(i));
204
+ seq.at(i) =
205
+ edit_distance2_backtracking(d, xs.at(i), ys.at(i), terminal_symbol);
206
+ }
207
+ return seq;
208
+ }
209
+
210
+ vector<vector<vector<uint32_t>>> suggested_ed2_path_with_delete(
211
+ vector<vector<uint32_t>>& xs,
212
+ vector<vector<uint32_t>>& ys,
213
+ uint32_t terminal_symbol,
214
+ uint32_t deletion_symbol) {
215
+ vector<vector<vector<uint32_t>>> seq(xs.size());
216
+ for (uint32_t i = 0; i < xs.size(); i++) {
217
+ vector<vector<uint32_t>> d = edit_distance2_with_dp(xs.at(i), ys.at(i));
218
+ seq.at(i) = edit_distance2_backtracking_with_delete(
219
+ d, xs.at(i), ys.at(i), terminal_symbol, deletion_symbol);
220
+ }
221
+ return seq;
222
+ }
223
+
224
+ PYBIND11_MODULE(libnat, m) {
225
+ m.def("compute_ed2", &compute_ed2, "compute_ed2");
226
+ m.def("suggested_ed2_path", &suggested_ed2_path, "suggested_ed2_path");
227
+ m.def(
228
+ "suggested_ed2_path_with_delete",
229
+ &suggested_ed2_path_with_delete,
230
+ "suggested_ed2_path_with_delete");
231
+ }
data/fairseq/fairseq/distributed/__pycache__/__init__.cpython-38.pyc ADDED
Binary file (673 Bytes). View file
 
data/fairseq/fairseq/distributed/__pycache__/fully_sharded_data_parallel.cpython-38.pyc ADDED
Binary file (4.84 kB). View file
 
data/fairseq/fairseq/distributed/__pycache__/module_proxy_wrapper.cpython-38.pyc ADDED
Binary file (2.18 kB). View file
 
data/fairseq/fairseq/distributed/__pycache__/tpu_distributed_data_parallel.cpython-38.pyc ADDED
Binary file (1.49 kB). View file
 
data/fairseq/fairseq/distributed/__pycache__/utils.cpython-38.pyc ADDED
Binary file (22.2 kB). View file
 
data/fairseq/fairseq/optim/__init__.py ADDED
@@ -0,0 +1,48 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Copyright (c) Facebook, Inc. and its affiliates.
2
+ #
3
+ # This source code is licensed under the MIT license found in the
4
+ # LICENSE file in the root directory of this source tree.
5
+ """isort:skip_file"""
6
+
7
+ import importlib
8
+ import os
9
+
10
+ from fairseq import registry
11
+ from fairseq.optim.bmuf import FairseqBMUF # noqa
12
+ from fairseq.optim.fairseq_optimizer import ( # noqa
13
+ FairseqOptimizer,
14
+ LegacyFairseqOptimizer,
15
+ )
16
+ from fairseq.optim.amp_optimizer import AMPOptimizer
17
+ from fairseq.optim.fp16_optimizer import FP16Optimizer, MemoryEfficientFP16Optimizer
18
+ from fairseq.optim.shard import shard_
19
+ from omegaconf import DictConfig
20
+
21
+ __all__ = [
22
+ "AMPOptimizer",
23
+ "FairseqOptimizer",
24
+ "FP16Optimizer",
25
+ "MemoryEfficientFP16Optimizer",
26
+ "shard_",
27
+ ]
28
+
29
+ (
30
+ _build_optimizer,
31
+ register_optimizer,
32
+ OPTIMIZER_REGISTRY,
33
+ OPTIMIZER_DATACLASS_REGISTRY,
34
+ ) = registry.setup_registry("--optimizer", base_class=FairseqOptimizer, required=True)
35
+
36
+
37
+ def build_optimizer(cfg: DictConfig, params, *extra_args, **extra_kwargs):
38
+ if all(isinstance(p, dict) for p in params):
39
+ params = [t for p in params for t in p.values()]
40
+ params = list(filter(lambda p: p.requires_grad, params))
41
+ return _build_optimizer(cfg, params, *extra_args, **extra_kwargs)
42
+
43
+
44
+ # automatically import any Python files in the optim/ directory
45
+ for file in sorted(os.listdir(os.path.dirname(__file__))):
46
+ if file.endswith(".py") and not file.startswith("_"):
47
+ file_name = file[: file.find(".py")]
48
+ importlib.import_module("fairseq.optim." + file_name)
data/fairseq/fairseq/optim/adadelta.py ADDED
@@ -0,0 +1,47 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Copyright (c) Facebook, Inc. and its affiliates.
2
+ #
3
+ # This source code is licensed under the MIT license found in the
4
+ # LICENSE file in the root directory of this source tree.
5
+
6
+ import torch.optim
7
+
8
+ from . import LegacyFairseqOptimizer, register_optimizer
9
+
10
+
11
+ @register_optimizer("adadelta")
12
+ class Adadelta(LegacyFairseqOptimizer):
13
+ def __init__(self, args, params):
14
+ super().__init__(args)
15
+ self._optimizer = torch.optim.Adadelta(params, **self.optimizer_config)
16
+
17
+ @staticmethod
18
+ def add_args(parser):
19
+ """Add optimizer-specific arguments to the parser."""
20
+ # fmt: off
21
+ parser.add_argument('--adadelta-rho', type=float, default=0.9, metavar='RHO',
22
+ help='coefficient used for computing a running average of squared gradients')
23
+ parser.add_argument('--adadelta-eps', type=float, default=1e-6, metavar='EPS',
24
+ help='term added to the denominator to improve numerical stability')
25
+ parser.add_argument('--weight-decay', '--wd', default=0.0, type=float, metavar='WD',
26
+ help='weight decay')
27
+ parser.add_argument('--anneal-eps', action='store_true', help='flag to anneal eps')
28
+ # fmt: on
29
+
30
+ @property
31
+ def optimizer_config(self):
32
+ """
33
+ Return a kwarg dictionary that will be used to override optimizer
34
+ args stored in checkpoints. This allows us to load a checkpoint and
35
+ resume training using a different set of optimizer args, e.g., with a
36
+ different learning rate.
37
+ """
38
+ return {
39
+ "lr": self.args.lr[0],
40
+ "rho": self.args.adadelta_rho,
41
+ "eps": self.args.adadelta_eps,
42
+ "weight_decay": self.args.weight_decay,
43
+ }
44
+
45
+ @property
46
+ def supports_flat_params(self):
47
+ return True
data/fairseq/fairseq/optim/adam.py ADDED
@@ -0,0 +1,239 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Copyright (c) Facebook, Inc. and its affiliates.
2
+ #
3
+ # This source code is licensed under the MIT license found in the
4
+ # LICENSE file in the root directory of this source tree.
5
+
6
+ import logging
7
+ import math
8
+ from collections.abc import Collection
9
+ from dataclasses import dataclass, field
10
+ from typing import Any, List
11
+
12
+ import torch
13
+ import torch.distributed as dist
14
+ import torch.optim
15
+ from fairseq.dataclass import FairseqDataclass
16
+ from fairseq.optim import FairseqOptimizer, register_optimizer
17
+ from fairseq.optim.fused_adam import get_fused_adam_class
18
+ from omegaconf import II, OmegaConf
19
+
20
+
21
+ logger = logging.getLogger(__name__)
22
+
23
+
24
+ @dataclass
25
+ class FairseqAdamConfig(FairseqDataclass):
26
+ adam_betas: Any = field(
27
+ default=(0.9, 0.999), metadata={"help": "betas for Adam optimizer"}
28
+ )
29
+ adam_eps: float = field(
30
+ default=1e-8, metadata={"help": "epsilon for Adam optimizer"}
31
+ )
32
+ weight_decay: float = field(default=0.0, metadata={"help": "weight decay"})
33
+ use_old_adam: bool = field(
34
+ default=False, metadata={"help": "Use fairseq.optim.adam.Adam"}
35
+ )
36
+ fp16_adam_stats: bool = field(
37
+ default=False, metadata={"help": "use FP16 stats (with automatic scaling)"}
38
+ )
39
+ # TODO common vars below in parent
40
+ tpu: bool = II("common.tpu")
41
+ lr: List[float] = II("optimization.lr")
42
+
43
+
44
+ @register_optimizer("adam", dataclass=FairseqAdamConfig)
45
+ class FairseqAdam(FairseqOptimizer):
46
+ """Adam optimizer for fairseq.
47
+
48
+ Important note: this optimizer corresponds to the "AdamW" variant of
49
+ Adam in its weight decay behavior. As such, it is most closely
50
+ analogous to torch.optim.AdamW from PyTorch.
51
+ """
52
+
53
+ def __init__(self, cfg: FairseqAdamConfig, params):
54
+ super().__init__(cfg)
55
+ fused_adam_cls = get_fused_adam_class()
56
+ use_fused_adam = (
57
+ not getattr(cfg, "use_old_adam", False)
58
+ and fused_adam_cls is not None
59
+ and torch.cuda.is_available()
60
+ )
61
+ if getattr(cfg, "tpu", False):
62
+ if self.cfg.fp16_adam_stats:
63
+ raise NotImplementedError("--fp16-adam-stats is only supported on GPU")
64
+ # on TPUs we use the Adam defined here, since it
65
+ # automatically casts gradients to FP32
66
+ self._optimizer = Adam(params, **self.optimizer_config)
67
+ elif use_fused_adam:
68
+ logger.info("using FusedAdam")
69
+ self._optimizer = fused_adam_cls(
70
+ params, use_fp16_stats=self.cfg.fp16_adam_stats, **self.optimizer_config
71
+ )
72
+ else:
73
+ if self.cfg.fp16_adam_stats:
74
+ raise NotImplementedError(
75
+ "--fp16-adam-stats is only supported with FusedAdamV1"
76
+ )
77
+ self._optimizer = Adam(params, **self.optimizer_config)
78
+
79
+ @property
80
+ def optimizer_config(self):
81
+ """
82
+ Return a kwarg dictionary that will be used to override optimizer
83
+ args stored in checkpoints. This allows us to load a checkpoint and
84
+ resume training using a different set of optimizer args, e.g., with a
85
+ different learning rate.
86
+ """
87
+ return {
88
+ "lr": self.cfg.lr[0]
89
+ if isinstance(self.cfg.lr, Collection)
90
+ else self.cfg.lr,
91
+ "betas": eval(self.cfg.adam_betas)
92
+ if isinstance(self.cfg.adam_betas, str)
93
+ else OmegaConf.to_container(self.cfg.adam_betas),
94
+ "eps": self.cfg.adam_eps,
95
+ "weight_decay": self.cfg.weight_decay,
96
+ }
97
+
98
+ def average_params(self):
99
+ """Reduce Params is only used during BMUF distributed training."""
100
+ state_dict = self.optimizer.state_dict()
101
+ total_gpus = float(dist.get_world_size())
102
+
103
+ for _, value in state_dict["state"].items():
104
+ value["exp_avg"] /= total_gpus
105
+ value["exp_avg_sq"] /= total_gpus
106
+ dist.all_reduce(value["exp_avg"], op=dist.ReduceOp.SUM)
107
+ dist.all_reduce(value["exp_avg_sq"], op=dist.ReduceOp.SUM)
108
+
109
+
110
+ class Adam(torch.optim.Optimizer):
111
+ r"""Implements Adam algorithm.
112
+
113
+ This implementation is modified from torch.optim.Adam based on:
114
+ `Fixed Weight Decay Regularization in Adam`
115
+ (see https://arxiv.org/abs/1711.05101)
116
+
117
+ It has been proposed in `Adam: A Method for Stochastic Optimization`_.
118
+
119
+ Args:
120
+ params (iterable): iterable of parameters to optimize or dicts defining
121
+ parameter groups
122
+ lr (float, optional): learning rate (default: 1e-3)
123
+ betas (Tuple[float, float], optional): coefficients used for computing
124
+ running averages of gradient and its square (default: (0.9, 0.999))
125
+ eps (float, optional): term added to the denominator to improve
126
+ numerical stability (default: 1e-8)
127
+ weight_decay (float, optional): weight decay (L2 penalty) (default: 0)
128
+ amsgrad (boolean, optional): whether to use the AMSGrad variant of this
129
+ algorithm from the paper `On the Convergence of Adam and Beyond`_
130
+
131
+ .. _Adam\: A Method for Stochastic Optimization:
132
+ https://arxiv.org/abs/1412.6980
133
+ .. _On the Convergence of Adam and Beyond:
134
+ https://openreview.net/forum?id=ryQu7f-RZ
135
+ """
136
+
137
+ def __init__(
138
+ self,
139
+ params,
140
+ lr=1e-3,
141
+ betas=(0.9, 0.999),
142
+ eps=1e-8,
143
+ weight_decay=0,
144
+ amsgrad=False,
145
+ ):
146
+ defaults = dict(
147
+ lr=lr, betas=betas, eps=eps, weight_decay=weight_decay, amsgrad=amsgrad
148
+ )
149
+ super(Adam, self).__init__(params, defaults)
150
+
151
+ @property
152
+ def supports_memory_efficient_fp16(self):
153
+ return True
154
+
155
+ @property
156
+ def supports_flat_params(self):
157
+ return True
158
+
159
+ def step(self, closure=None):
160
+ """Performs a single optimization step.
161
+
162
+ Args:
163
+ closure (callable, optional): A closure that reevaluates the model
164
+ and returns the loss.
165
+ """
166
+ loss = None
167
+ if closure is not None:
168
+ loss = closure()
169
+
170
+ for group in self.param_groups:
171
+ for p in group["params"]:
172
+ if p.grad is None:
173
+ continue
174
+ grad = p.grad.data
175
+ if grad.dtype in {torch.float16, torch.bfloat16}:
176
+ grad = grad.float()
177
+ if grad.is_sparse:
178
+ raise RuntimeError(
179
+ "Adam does not support sparse gradients, please consider SparseAdam instead"
180
+ )
181
+ amsgrad = group.get("amsgrad", False)
182
+
183
+ p_data_fp32 = p.data
184
+ if p.data.dtype in {torch.float16, torch.bfloat16}:
185
+ p_data_fp32 = p_data_fp32.float()
186
+
187
+ state = self.state[p]
188
+
189
+ # State initialization
190
+ if len(state) == 0:
191
+ state["step"] = 0
192
+ # Exponential moving average of gradient values
193
+ state["exp_avg"] = torch.zeros_like(p_data_fp32)
194
+ # Exponential moving average of squared gradient values
195
+ state["exp_avg_sq"] = torch.zeros_like(p_data_fp32)
196
+ if amsgrad:
197
+ # Maintains max of all exp. moving avg. of sq. grad. values
198
+ state["max_exp_avg_sq"] = torch.zeros_like(p_data_fp32)
199
+ else:
200
+ state["exp_avg"] = state["exp_avg"].to(p_data_fp32)
201
+ state["exp_avg_sq"] = state["exp_avg_sq"].to(p_data_fp32)
202
+ if amsgrad:
203
+ state["max_exp_avg_sq"] = state["max_exp_avg_sq"].to(
204
+ p_data_fp32
205
+ )
206
+
207
+ exp_avg, exp_avg_sq = state["exp_avg"], state["exp_avg_sq"]
208
+ if amsgrad:
209
+ max_exp_avg_sq = state["max_exp_avg_sq"]
210
+ beta1, beta2 = group["betas"]
211
+
212
+ state["step"] += 1
213
+
214
+ # Decay the first and second moment running average coefficient
215
+ exp_avg.mul_(beta1).add_(grad, alpha=1 - beta1)
216
+ exp_avg_sq.mul_(beta2).addcmul_(grad, grad, value=1 - beta2)
217
+ if amsgrad:
218
+ # Maintains the maximum of all 2nd moment running avg. till now
219
+ torch.max(max_exp_avg_sq, exp_avg_sq, out=max_exp_avg_sq)
220
+ # Use the max. for normalizing running avg. of gradient
221
+ denom = max_exp_avg_sq.sqrt().add_(group["eps"])
222
+ else:
223
+ denom = exp_avg_sq.sqrt().add_(group["eps"])
224
+
225
+ bias_correction1 = 1 - beta1 ** state["step"]
226
+ bias_correction2 = 1 - beta2 ** state["step"]
227
+ step_size = group["lr"] * math.sqrt(bias_correction2) / bias_correction1
228
+
229
+ if group["weight_decay"] != 0:
230
+ p_data_fp32.add_(
231
+ p_data_fp32, alpha=-group["weight_decay"] * group["lr"]
232
+ )
233
+
234
+ p_data_fp32.addcdiv_(exp_avg, denom, value=-step_size)
235
+
236
+ if p.data.dtype in {torch.float16, torch.bfloat16}:
237
+ p.data.copy_(p_data_fp32)
238
+
239
+ return loss
data/fairseq/fairseq/optim/bmuf.py ADDED
@@ -0,0 +1,200 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Copyright (c) Facebook, Inc. and its affiliates.
2
+ #
3
+ # This source code is licensed under the MIT license found in the
4
+ # LICENSE file in the root directory of this source tree.
5
+
6
+ from dataclasses import dataclass, field
7
+
8
+ import torch
9
+ import torch.distributed as dist
10
+ from fairseq.dataclass.configs import FairseqBMUFConfig
11
+ from fairseq.dataclass.utils import gen_parser_from_dataclass
12
+ from fairseq.optim.fairseq_optimizer import FairseqOptimizer
13
+
14
+
15
+ class FairseqBMUF(FairseqOptimizer):
16
+ """
17
+ Implements incremental block distributed data parallelism similar to
18
+ https://ieeexplore.ieee.org/document/7472805
19
+
20
+ Paper title: Scalable training of deep learning machines by incremental
21
+ block training with intra-block parallel optimization and blockwise
22
+ model-update filtering
23
+ """
24
+
25
+ def __init__(self, cfg: FairseqBMUFConfig, optimizer):
26
+ super().__init__(cfg)
27
+ self._optimizer = optimizer
28
+ self._num_updates = 0
29
+ self.sync_iter = cfg.global_sync_iter
30
+ self.block_momentum = cfg.block_momentum
31
+ self.block_lr = cfg.block_lr
32
+ self._reset_local_data()
33
+ self.warmup_iteration = cfg.warmup_iterations
34
+ self.use_nbm = cfg.use_nbm
35
+ self.initial_state = self._optimizer.state_dict()
36
+ self.average_sync = self.cfg.average_sync
37
+ self.world_size = self.cfg.distributed_world_size
38
+
39
+ @staticmethod
40
+ def add_args(parser):
41
+ """Add optimizer-specific arguments to the parser."""
42
+ gen_parser_from_dataclass(parser, FairseqBMUFConfig())
43
+
44
+ @property
45
+ def optimizer(self):
46
+ return self._optimizer.optimizer
47
+
48
+ @property
49
+ def optimizer_config(self):
50
+ return self._optimizer.optimizer_config
51
+
52
+ def get_lr(self):
53
+ return self._optimizer.get_lr()
54
+
55
+ def set_lr(self, lr):
56
+ self._optimizer.set_lr(lr)
57
+
58
+ def state_dict(self):
59
+ return self._optimizer.state_dict()
60
+
61
+ def load_state_dict(self, state_dict, optimizer_overrides=None):
62
+ self._optimizer.load_state_dict(state_dict, optimizer_overrides)
63
+ self.initial_state = self._optimizer.state_dict()
64
+
65
+ def multiply_grads(self, c):
66
+ """Multiplies grads by a constant *c*."""
67
+ self._optimizer.multiply_grads(c)
68
+
69
+ def clip_grad_norm(self, max_norm, aggregate_norm_fn=None):
70
+ """Clips gradient norm."""
71
+ return self._optimizer.clip_grad_norm(max_norm, aggregate_norm_fn)
72
+
73
+ def average_params(self):
74
+ self._optimizer.average_params()
75
+
76
+ def _block_sync(self):
77
+ if self.world_size <= 1:
78
+ return
79
+ # Update the global model using local models from all GPUs
80
+ # (Step-1) Calculate grad between previously synced model and
81
+ # currrent local model
82
+ if self.block_momentum != 0:
83
+ self._calc_grad()
84
+
85
+ # (Step-2) Average gradient from all GPUs
86
+ self._avg_grad_from_all_gpus()
87
+
88
+ # (Step-3) Calculate global momentum and update the global model
89
+ if self.block_momentum != 0:
90
+ self._update_global_model()
91
+
92
+ # (Step-4) Average local optimizer params
93
+ if self.average_sync:
94
+ self.average_params()
95
+
96
+ def _is_warmup_end(self):
97
+ # Check whether train iterations is equal to warmup iter
98
+ if self.get_num_updates() == self.warmup_iteration:
99
+ return True
100
+ return False
101
+
102
+ def _is_bmuf_iter(self):
103
+ # Check whether train iterations is equal to bmuf sync iter
104
+ if (self.get_num_updates() > self.warmup_iteration) and (
105
+ self.get_num_updates() % self.sync_iter == 0
106
+ ):
107
+ return True
108
+ return False
109
+
110
+ def _warmup_sync(self, root_rank=0):
111
+ if self.world_size <= 1:
112
+ return
113
+ # Broadcast the local model to all gpus
114
+ for param in self.params:
115
+ dist.broadcast(param.data, src=root_rank)
116
+
117
+ # Update local optimizer state
118
+ if self.average_sync:
119
+ self._optimizer.average_params()
120
+ else:
121
+ self._optimizer.load_state_dict(self.initial_state)
122
+
123
+ self._reset_local_data()
124
+
125
+ def step(self, closure=None):
126
+ """Performs a single optimization step."""
127
+ self._optimizer.step(closure)
128
+ self.set_num_updates(self.get_num_updates() + 1)
129
+ if self._is_warmup_end():
130
+ self._warmup_sync()
131
+ elif self._is_bmuf_iter():
132
+ self._block_sync()
133
+
134
+ def zero_grad(self):
135
+ """Clears the gradients of all optimized parameters."""
136
+ self._optimizer.zero_grad()
137
+
138
+ def get_num_updates(self):
139
+ """Get the number of parameters updates."""
140
+ return self._num_updates
141
+
142
+ def set_num_updates(self, num_updates):
143
+ """Set the number of parameters updates."""
144
+ self._num_updates = num_updates
145
+
146
+ @torch.no_grad()
147
+ def _reset_local_data(self):
148
+ # (Step-0) Initialize global momentum parameters and store global copy on each gpu
149
+ self.global_params = [torch.zeros_like(p.data) for p in self.params]
150
+ self.smoothed_grads = [p.data.new_zeros(p.data.size()) for p in self.params]
151
+ self.grads = [p.data.new_zeros(p.data.size()) for p in self.params]
152
+
153
+ # saving the global model locally for calculating gradient during bmuf sync
154
+ for param, global_param in zip(self.params, self.global_params):
155
+ global_param.copy_(param.data)
156
+
157
+ @torch.no_grad()
158
+ def _calc_grad(self):
159
+ # global_params is basically the global copy from the previously finished
160
+ # synchronisation. param.data is local parameter after block_sync_freq
161
+ # for the local gpu. so grad is difference between previously synced
162
+ # model and currrent local model.
163
+ for index, (param, global_param) in enumerate(
164
+ zip(self.params, self.global_params)
165
+ ):
166
+ self.grads[index] = global_param - param.data
167
+
168
+ def _avg_grad_from_all_gpus(self):
169
+ for index, param in enumerate(self.params):
170
+ sync_para = param.data if self.block_momentum == 0 else self.grads[index]
171
+ sync_para /= float(dist.get_world_size())
172
+ dist.all_reduce(sync_para, op=dist.ReduceOp.SUM)
173
+
174
+ @torch.no_grad()
175
+ def _update_global_model(self):
176
+ for index, (param, global_param, smoothed_grad, grad) in enumerate(
177
+ zip(
178
+ self.params,
179
+ self.global_params,
180
+ self.smoothed_grads,
181
+ # all gpus would share the same value of smoothed_grad, since it is
182
+ # always computed on synchronized gradients.
183
+ self.grads,
184
+ )
185
+ ):
186
+ # global_param is basically last syncrhornized parameter. though
187
+ # smoothed_grad is local, all processes will have same value of
188
+ # smoothed_grad and hence param is globally synchronized copy.
189
+ # smoothed_grad(t) = BM * smoothed_grad(t-1) + BM_lr * grad(t)
190
+ smoothed_grad = self.block_momentum * smoothed_grad + self.block_lr * grad
191
+ param.data.copy_(global_param - smoothed_grad)
192
+
193
+ # A Nesterov momentum here is to do a partial weight update before
194
+ # calculating the gradient
195
+ if self.use_nbm:
196
+ param.data.copy_(param.data - self.block_momentum * smoothed_grad)
197
+
198
+ # backup for the next synchronization.
199
+ self.smoothed_grads[index] = smoothed_grad
200
+ global_param.copy_(param.data)
data/train.eng_Latn ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b8fc7a8095a8dca8cb0b0582c3ebe293486bef635234a0968ce4b3dcf8a50268
3
+ size 16731703
data/train.mni_Beng ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1efcc7e8a09ba7cef6cf115342ae16d34058b38da1b965984dee7090df8536cf
3
+ size 43201990
nllb-mni-lsftl/checkpoint-26112/README.md ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: facebook/nllb-200-distilled-600M
3
+ library_name: peft
4
+ ---
5
+
6
+ # Model Card for Model ID
7
+
8
+ <!-- Provide a quick summary of what the model is/does. -->
9
+
10
+
11
+
12
+ ## Model Details
13
+
14
+ ### Model Description
15
+
16
+ <!-- Provide a longer summary of what this model is. -->
17
+
18
+
19
+
20
+ - **Developed by:** [More Information Needed]
21
+ - **Funded by [optional]:** [More Information Needed]
22
+ - **Shared by [optional]:** [More Information Needed]
23
+ - **Model type:** [More Information Needed]
24
+ - **Language(s) (NLP):** [More Information Needed]
25
+ - **License:** [More Information Needed]
26
+ - **Finetuned from model [optional]:** [More Information Needed]
27
+
28
+ ### Model Sources [optional]
29
+
30
+ <!-- Provide the basic links for the model. -->
31
+
32
+ - **Repository:** [More Information Needed]
33
+ - **Paper [optional]:** [More Information Needed]
34
+ - **Demo [optional]:** [More Information Needed]
35
+
36
+ ## Uses
37
+
38
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
39
+
40
+ ### Direct Use
41
+
42
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
43
+
44
+ [More Information Needed]
45
+
46
+ ### Downstream Use [optional]
47
+
48
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
49
+
50
+ [More Information Needed]
51
+
52
+ ### Out-of-Scope Use
53
+
54
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
55
+
56
+ [More Information Needed]
57
+
58
+ ## Bias, Risks, and Limitations
59
+
60
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
61
+
62
+ [More Information Needed]
63
+
64
+ ### Recommendations
65
+
66
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
67
+
68
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
69
+
70
+ ## How to Get Started with the Model
71
+
72
+ Use the code below to get started with the model.
73
+
74
+ [More Information Needed]
75
+
76
+ ## Training Details
77
+
78
+ ### Training Data
79
+
80
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
81
+
82
+ [More Information Needed]
83
+
84
+ ### Training Procedure
85
+
86
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
87
+
88
+ #### Preprocessing [optional]
89
+
90
+ [More Information Needed]
91
+
92
+
93
+ #### Training Hyperparameters
94
+
95
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
96
+
97
+ #### Speeds, Sizes, Times [optional]
98
+
99
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
100
+
101
+ [More Information Needed]
102
+
103
+ ## Evaluation
104
+
105
+ <!-- This section describes the evaluation protocols and provides the results. -->
106
+
107
+ ### Testing Data, Factors & Metrics
108
+
109
+ #### Testing Data
110
+
111
+ <!-- This should link to a Dataset Card if possible. -->
112
+
113
+ [More Information Needed]
114
+
115
+ #### Factors
116
+
117
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
118
+
119
+ [More Information Needed]
120
+
121
+ #### Metrics
122
+
123
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
124
+
125
+ [More Information Needed]
126
+
127
+ ### Results
128
+
129
+ [More Information Needed]
130
+
131
+ #### Summary
132
+
133
+
134
+
135
+ ## Model Examination [optional]
136
+
137
+ <!-- Relevant interpretability work for the model goes here -->
138
+
139
+ [More Information Needed]
140
+
141
+ ## Environmental Impact
142
+
143
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
144
+
145
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
146
+
147
+ - **Hardware Type:** [More Information Needed]
148
+ - **Hours used:** [More Information Needed]
149
+ - **Cloud Provider:** [More Information Needed]
150
+ - **Compute Region:** [More Information Needed]
151
+ - **Carbon Emitted:** [More Information Needed]
152
+
153
+ ## Technical Specifications [optional]
154
+
155
+ ### Model Architecture and Objective
156
+
157
+ [More Information Needed]
158
+
159
+ ### Compute Infrastructure
160
+
161
+ [More Information Needed]
162
+
163
+ #### Hardware
164
+
165
+ [More Information Needed]
166
+
167
+ #### Software
168
+
169
+ [More Information Needed]
170
+
171
+ ## Citation [optional]
172
+
173
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
174
+
175
+ **BibTeX:**
176
+
177
+ [More Information Needed]
178
+
179
+ **APA:**
180
+
181
+ [More Information Needed]
182
+
183
+ ## Glossary [optional]
184
+
185
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
186
+
187
+ [More Information Needed]
188
+
189
+ ## More Information [optional]
190
+
191
+ [More Information Needed]
192
+
193
+ ## Model Card Authors [optional]
194
+
195
+ [More Information Needed]
196
+
197
+ ## Model Card Contact
198
+
199
+ [More Information Needed]
200
+ ### Framework versions
201
+
202
+ - PEFT 0.13.2
nllb-mni-lsftl/checkpoint-26112/adapter_config.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "facebook/nllb-200-distilled-600M",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layer_replication": null,
10
+ "layers_pattern": null,
11
+ "layers_to_transform": null,
12
+ "loftq_config": {},
13
+ "lora_alpha": 32,
14
+ "lora_dropout": 0.05,
15
+ "megatron_config": null,
16
+ "megatron_core": "megatron.core",
17
+ "modules_to_save": null,
18
+ "peft_type": "LORA",
19
+ "r": 16,
20
+ "rank_pattern": {},
21
+ "revision": null,
22
+ "target_modules": [
23
+ "q_proj",
24
+ "v_proj",
25
+ "out_proj",
26
+ "k_proj"
27
+ ],
28
+ "task_type": "SEQ_2_SEQ_LM",
29
+ "use_dora": false,
30
+ "use_rslora": false
31
+ }
nllb-mni-lsftl/checkpoint-26112/special_tokens_map.json ADDED
@@ -0,0 +1,255 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "ace_Arab",
4
+ "ace_Latn",
5
+ "acm_Arab",
6
+ "acq_Arab",
7
+ "aeb_Arab",
8
+ "afr_Latn",
9
+ "ajp_Arab",
10
+ "aka_Latn",
11
+ "amh_Ethi",
12
+ "apc_Arab",
13
+ "arb_Arab",
14
+ "ars_Arab",
15
+ "ary_Arab",
16
+ "arz_Arab",
17
+ "asm_Beng",
18
+ "ast_Latn",
19
+ "awa_Deva",
20
+ "ayr_Latn",
21
+ "azb_Arab",
22
+ "azj_Latn",
23
+ "bak_Cyrl",
24
+ "bam_Latn",
25
+ "ban_Latn",
26
+ "bel_Cyrl",
27
+ "bem_Latn",
28
+ "ben_Beng",
29
+ "bho_Deva",
30
+ "bjn_Arab",
31
+ "bjn_Latn",
32
+ "bod_Tibt",
33
+ "bos_Latn",
34
+ "bug_Latn",
35
+ "bul_Cyrl",
36
+ "cat_Latn",
37
+ "ceb_Latn",
38
+ "ces_Latn",
39
+ "cjk_Latn",
40
+ "ckb_Arab",
41
+ "crh_Latn",
42
+ "cym_Latn",
43
+ "dan_Latn",
44
+ "deu_Latn",
45
+ "dik_Latn",
46
+ "dyu_Latn",
47
+ "dzo_Tibt",
48
+ "ell_Grek",
49
+ "eng_Latn",
50
+ "epo_Latn",
51
+ "est_Latn",
52
+ "eus_Latn",
53
+ "ewe_Latn",
54
+ "fao_Latn",
55
+ "pes_Arab",
56
+ "fij_Latn",
57
+ "fin_Latn",
58
+ "fon_Latn",
59
+ "fra_Latn",
60
+ "fur_Latn",
61
+ "fuv_Latn",
62
+ "gla_Latn",
63
+ "gle_Latn",
64
+ "glg_Latn",
65
+ "grn_Latn",
66
+ "guj_Gujr",
67
+ "hat_Latn",
68
+ "hau_Latn",
69
+ "heb_Hebr",
70
+ "hin_Deva",
71
+ "hne_Deva",
72
+ "hrv_Latn",
73
+ "hun_Latn",
74
+ "hye_Armn",
75
+ "ibo_Latn",
76
+ "ilo_Latn",
77
+ "ind_Latn",
78
+ "isl_Latn",
79
+ "ita_Latn",
80
+ "jav_Latn",
81
+ "jpn_Jpan",
82
+ "kab_Latn",
83
+ "kac_Latn",
84
+ "kam_Latn",
85
+ "kan_Knda",
86
+ "kas_Arab",
87
+ "kas_Deva",
88
+ "kat_Geor",
89
+ "knc_Arab",
90
+ "knc_Latn",
91
+ "kaz_Cyrl",
92
+ "kbp_Latn",
93
+ "kea_Latn",
94
+ "khm_Khmr",
95
+ "kik_Latn",
96
+ "kin_Latn",
97
+ "kir_Cyrl",
98
+ "kmb_Latn",
99
+ "kon_Latn",
100
+ "kor_Hang",
101
+ "kmr_Latn",
102
+ "lao_Laoo",
103
+ "lvs_Latn",
104
+ "lij_Latn",
105
+ "lim_Latn",
106
+ "lin_Latn",
107
+ "lit_Latn",
108
+ "lmo_Latn",
109
+ "ltg_Latn",
110
+ "ltz_Latn",
111
+ "lua_Latn",
112
+ "lug_Latn",
113
+ "luo_Latn",
114
+ "lus_Latn",
115
+ "mag_Deva",
116
+ "mai_Deva",
117
+ "mal_Mlym",
118
+ "mar_Deva",
119
+ "min_Latn",
120
+ "mkd_Cyrl",
121
+ "plt_Latn",
122
+ "mlt_Latn",
123
+ "mni_Beng",
124
+ "khk_Cyrl",
125
+ "mos_Latn",
126
+ "mri_Latn",
127
+ "zsm_Latn",
128
+ "mya_Mymr",
129
+ "nld_Latn",
130
+ "nno_Latn",
131
+ "nob_Latn",
132
+ "npi_Deva",
133
+ "nso_Latn",
134
+ "nus_Latn",
135
+ "nya_Latn",
136
+ "oci_Latn",
137
+ "gaz_Latn",
138
+ "ory_Orya",
139
+ "pag_Latn",
140
+ "pan_Guru",
141
+ "pap_Latn",
142
+ "pol_Latn",
143
+ "por_Latn",
144
+ "prs_Arab",
145
+ "pbt_Arab",
146
+ "quy_Latn",
147
+ "ron_Latn",
148
+ "run_Latn",
149
+ "rus_Cyrl",
150
+ "sag_Latn",
151
+ "san_Deva",
152
+ "sat_Beng",
153
+ "scn_Latn",
154
+ "shn_Mymr",
155
+ "sin_Sinh",
156
+ "slk_Latn",
157
+ "slv_Latn",
158
+ "smo_Latn",
159
+ "sna_Latn",
160
+ "snd_Arab",
161
+ "som_Latn",
162
+ "sot_Latn",
163
+ "spa_Latn",
164
+ "als_Latn",
165
+ "srd_Latn",
166
+ "srp_Cyrl",
167
+ "ssw_Latn",
168
+ "sun_Latn",
169
+ "swe_Latn",
170
+ "swh_Latn",
171
+ "szl_Latn",
172
+ "tam_Taml",
173
+ "tat_Cyrl",
174
+ "tel_Telu",
175
+ "tgk_Cyrl",
176
+ "tgl_Latn",
177
+ "tha_Thai",
178
+ "tir_Ethi",
179
+ "taq_Latn",
180
+ "taq_Tfng",
181
+ "tpi_Latn",
182
+ "tsn_Latn",
183
+ "tso_Latn",
184
+ "tuk_Latn",
185
+ "tum_Latn",
186
+ "tur_Latn",
187
+ "twi_Latn",
188
+ "tzm_Tfng",
189
+ "uig_Arab",
190
+ "ukr_Cyrl",
191
+ "umb_Latn",
192
+ "urd_Arab",
193
+ "uzn_Latn",
194
+ "vec_Latn",
195
+ "vie_Latn",
196
+ "war_Latn",
197
+ "wol_Latn",
198
+ "xho_Latn",
199
+ "ydd_Hebr",
200
+ "yor_Latn",
201
+ "yue_Hant",
202
+ "zho_Hans",
203
+ "zho_Hant",
204
+ "zul_Latn"
205
+ ],
206
+ "bos_token": {
207
+ "content": "<s>",
208
+ "lstrip": false,
209
+ "normalized": false,
210
+ "rstrip": false,
211
+ "single_word": false
212
+ },
213
+ "cls_token": {
214
+ "content": "<s>",
215
+ "lstrip": false,
216
+ "normalized": false,
217
+ "rstrip": false,
218
+ "single_word": false
219
+ },
220
+ "eos_token": {
221
+ "content": "</s>",
222
+ "lstrip": false,
223
+ "normalized": false,
224
+ "rstrip": false,
225
+ "single_word": false
226
+ },
227
+ "mask_token": {
228
+ "content": "<mask>",
229
+ "lstrip": true,
230
+ "normalized": true,
231
+ "rstrip": false,
232
+ "single_word": false
233
+ },
234
+ "pad_token": {
235
+ "content": "<pad>",
236
+ "lstrip": false,
237
+ "normalized": false,
238
+ "rstrip": false,
239
+ "single_word": false
240
+ },
241
+ "sep_token": {
242
+ "content": "</s>",
243
+ "lstrip": false,
244
+ "normalized": false,
245
+ "rstrip": false,
246
+ "single_word": false
247
+ },
248
+ "unk_token": {
249
+ "content": "<unk>",
250
+ "lstrip": false,
251
+ "normalized": false,
252
+ "rstrip": false,
253
+ "single_word": false
254
+ }
255
+ }
nllb-mni-lsftl/checkpoint-26112/tokenizer_config.json ADDED
@@ -0,0 +1,1878 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "256001": {
36
+ "content": "ace_Arab",
37
+ "lstrip": false,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ },
43
+ "256002": {
44
+ "content": "ace_Latn",
45
+ "lstrip": false,
46
+ "normalized": false,
47
+ "rstrip": false,
48
+ "single_word": false,
49
+ "special": true
50
+ },
51
+ "256003": {
52
+ "content": "acm_Arab",
53
+ "lstrip": false,
54
+ "normalized": false,
55
+ "rstrip": false,
56
+ "single_word": false,
57
+ "special": true
58
+ },
59
+ "256004": {
60
+ "content": "acq_Arab",
61
+ "lstrip": false,
62
+ "normalized": false,
63
+ "rstrip": false,
64
+ "single_word": false,
65
+ "special": true
66
+ },
67
+ "256005": {
68
+ "content": "aeb_Arab",
69
+ "lstrip": false,
70
+ "normalized": false,
71
+ "rstrip": false,
72
+ "single_word": false,
73
+ "special": true
74
+ },
75
+ "256006": {
76
+ "content": "afr_Latn",
77
+ "lstrip": false,
78
+ "normalized": false,
79
+ "rstrip": false,
80
+ "single_word": false,
81
+ "special": true
82
+ },
83
+ "256007": {
84
+ "content": "ajp_Arab",
85
+ "lstrip": false,
86
+ "normalized": false,
87
+ "rstrip": false,
88
+ "single_word": false,
89
+ "special": true
90
+ },
91
+ "256008": {
92
+ "content": "aka_Latn",
93
+ "lstrip": false,
94
+ "normalized": false,
95
+ "rstrip": false,
96
+ "single_word": false,
97
+ "special": true
98
+ },
99
+ "256009": {
100
+ "content": "amh_Ethi",
101
+ "lstrip": false,
102
+ "normalized": false,
103
+ "rstrip": false,
104
+ "single_word": false,
105
+ "special": true
106
+ },
107
+ "256010": {
108
+ "content": "apc_Arab",
109
+ "lstrip": false,
110
+ "normalized": false,
111
+ "rstrip": false,
112
+ "single_word": false,
113
+ "special": true
114
+ },
115
+ "256011": {
116
+ "content": "arb_Arab",
117
+ "lstrip": false,
118
+ "normalized": false,
119
+ "rstrip": false,
120
+ "single_word": false,
121
+ "special": true
122
+ },
123
+ "256012": {
124
+ "content": "ars_Arab",
125
+ "lstrip": false,
126
+ "normalized": false,
127
+ "rstrip": false,
128
+ "single_word": false,
129
+ "special": true
130
+ },
131
+ "256013": {
132
+ "content": "ary_Arab",
133
+ "lstrip": false,
134
+ "normalized": false,
135
+ "rstrip": false,
136
+ "single_word": false,
137
+ "special": true
138
+ },
139
+ "256014": {
140
+ "content": "arz_Arab",
141
+ "lstrip": false,
142
+ "normalized": false,
143
+ "rstrip": false,
144
+ "single_word": false,
145
+ "special": true
146
+ },
147
+ "256015": {
148
+ "content": "asm_Beng",
149
+ "lstrip": false,
150
+ "normalized": false,
151
+ "rstrip": false,
152
+ "single_word": false,
153
+ "special": true
154
+ },
155
+ "256016": {
156
+ "content": "ast_Latn",
157
+ "lstrip": false,
158
+ "normalized": false,
159
+ "rstrip": false,
160
+ "single_word": false,
161
+ "special": true
162
+ },
163
+ "256017": {
164
+ "content": "awa_Deva",
165
+ "lstrip": false,
166
+ "normalized": false,
167
+ "rstrip": false,
168
+ "single_word": false,
169
+ "special": true
170
+ },
171
+ "256018": {
172
+ "content": "ayr_Latn",
173
+ "lstrip": false,
174
+ "normalized": false,
175
+ "rstrip": false,
176
+ "single_word": false,
177
+ "special": true
178
+ },
179
+ "256019": {
180
+ "content": "azb_Arab",
181
+ "lstrip": false,
182
+ "normalized": false,
183
+ "rstrip": false,
184
+ "single_word": false,
185
+ "special": true
186
+ },
187
+ "256020": {
188
+ "content": "azj_Latn",
189
+ "lstrip": false,
190
+ "normalized": false,
191
+ "rstrip": false,
192
+ "single_word": false,
193
+ "special": true
194
+ },
195
+ "256021": {
196
+ "content": "bak_Cyrl",
197
+ "lstrip": false,
198
+ "normalized": false,
199
+ "rstrip": false,
200
+ "single_word": false,
201
+ "special": true
202
+ },
203
+ "256022": {
204
+ "content": "bam_Latn",
205
+ "lstrip": false,
206
+ "normalized": false,
207
+ "rstrip": false,
208
+ "single_word": false,
209
+ "special": true
210
+ },
211
+ "256023": {
212
+ "content": "ban_Latn",
213
+ "lstrip": false,
214
+ "normalized": false,
215
+ "rstrip": false,
216
+ "single_word": false,
217
+ "special": true
218
+ },
219
+ "256024": {
220
+ "content": "bel_Cyrl",
221
+ "lstrip": false,
222
+ "normalized": false,
223
+ "rstrip": false,
224
+ "single_word": false,
225
+ "special": true
226
+ },
227
+ "256025": {
228
+ "content": "bem_Latn",
229
+ "lstrip": false,
230
+ "normalized": false,
231
+ "rstrip": false,
232
+ "single_word": false,
233
+ "special": true
234
+ },
235
+ "256026": {
236
+ "content": "ben_Beng",
237
+ "lstrip": false,
238
+ "normalized": false,
239
+ "rstrip": false,
240
+ "single_word": false,
241
+ "special": true
242
+ },
243
+ "256027": {
244
+ "content": "bho_Deva",
245
+ "lstrip": false,
246
+ "normalized": false,
247
+ "rstrip": false,
248
+ "single_word": false,
249
+ "special": true
250
+ },
251
+ "256028": {
252
+ "content": "bjn_Arab",
253
+ "lstrip": false,
254
+ "normalized": false,
255
+ "rstrip": false,
256
+ "single_word": false,
257
+ "special": true
258
+ },
259
+ "256029": {
260
+ "content": "bjn_Latn",
261
+ "lstrip": false,
262
+ "normalized": false,
263
+ "rstrip": false,
264
+ "single_word": false,
265
+ "special": true
266
+ },
267
+ "256030": {
268
+ "content": "bod_Tibt",
269
+ "lstrip": false,
270
+ "normalized": false,
271
+ "rstrip": false,
272
+ "single_word": false,
273
+ "special": true
274
+ },
275
+ "256031": {
276
+ "content": "bos_Latn",
277
+ "lstrip": false,
278
+ "normalized": false,
279
+ "rstrip": false,
280
+ "single_word": false,
281
+ "special": true
282
+ },
283
+ "256032": {
284
+ "content": "bug_Latn",
285
+ "lstrip": false,
286
+ "normalized": false,
287
+ "rstrip": false,
288
+ "single_word": false,
289
+ "special": true
290
+ },
291
+ "256033": {
292
+ "content": "bul_Cyrl",
293
+ "lstrip": false,
294
+ "normalized": false,
295
+ "rstrip": false,
296
+ "single_word": false,
297
+ "special": true
298
+ },
299
+ "256034": {
300
+ "content": "cat_Latn",
301
+ "lstrip": false,
302
+ "normalized": false,
303
+ "rstrip": false,
304
+ "single_word": false,
305
+ "special": true
306
+ },
307
+ "256035": {
308
+ "content": "ceb_Latn",
309
+ "lstrip": false,
310
+ "normalized": false,
311
+ "rstrip": false,
312
+ "single_word": false,
313
+ "special": true
314
+ },
315
+ "256036": {
316
+ "content": "ces_Latn",
317
+ "lstrip": false,
318
+ "normalized": false,
319
+ "rstrip": false,
320
+ "single_word": false,
321
+ "special": true
322
+ },
323
+ "256037": {
324
+ "content": "cjk_Latn",
325
+ "lstrip": false,
326
+ "normalized": false,
327
+ "rstrip": false,
328
+ "single_word": false,
329
+ "special": true
330
+ },
331
+ "256038": {
332
+ "content": "ckb_Arab",
333
+ "lstrip": false,
334
+ "normalized": false,
335
+ "rstrip": false,
336
+ "single_word": false,
337
+ "special": true
338
+ },
339
+ "256039": {
340
+ "content": "crh_Latn",
341
+ "lstrip": false,
342
+ "normalized": false,
343
+ "rstrip": false,
344
+ "single_word": false,
345
+ "special": true
346
+ },
347
+ "256040": {
348
+ "content": "cym_Latn",
349
+ "lstrip": false,
350
+ "normalized": false,
351
+ "rstrip": false,
352
+ "single_word": false,
353
+ "special": true
354
+ },
355
+ "256041": {
356
+ "content": "dan_Latn",
357
+ "lstrip": false,
358
+ "normalized": false,
359
+ "rstrip": false,
360
+ "single_word": false,
361
+ "special": true
362
+ },
363
+ "256042": {
364
+ "content": "deu_Latn",
365
+ "lstrip": false,
366
+ "normalized": false,
367
+ "rstrip": false,
368
+ "single_word": false,
369
+ "special": true
370
+ },
371
+ "256043": {
372
+ "content": "dik_Latn",
373
+ "lstrip": false,
374
+ "normalized": false,
375
+ "rstrip": false,
376
+ "single_word": false,
377
+ "special": true
378
+ },
379
+ "256044": {
380
+ "content": "dyu_Latn",
381
+ "lstrip": false,
382
+ "normalized": false,
383
+ "rstrip": false,
384
+ "single_word": false,
385
+ "special": true
386
+ },
387
+ "256045": {
388
+ "content": "dzo_Tibt",
389
+ "lstrip": false,
390
+ "normalized": false,
391
+ "rstrip": false,
392
+ "single_word": false,
393
+ "special": true
394
+ },
395
+ "256046": {
396
+ "content": "ell_Grek",
397
+ "lstrip": false,
398
+ "normalized": false,
399
+ "rstrip": false,
400
+ "single_word": false,
401
+ "special": true
402
+ },
403
+ "256047": {
404
+ "content": "eng_Latn",
405
+ "lstrip": false,
406
+ "normalized": false,
407
+ "rstrip": false,
408
+ "single_word": false,
409
+ "special": true
410
+ },
411
+ "256048": {
412
+ "content": "epo_Latn",
413
+ "lstrip": false,
414
+ "normalized": false,
415
+ "rstrip": false,
416
+ "single_word": false,
417
+ "special": true
418
+ },
419
+ "256049": {
420
+ "content": "est_Latn",
421
+ "lstrip": false,
422
+ "normalized": false,
423
+ "rstrip": false,
424
+ "single_word": false,
425
+ "special": true
426
+ },
427
+ "256050": {
428
+ "content": "eus_Latn",
429
+ "lstrip": false,
430
+ "normalized": false,
431
+ "rstrip": false,
432
+ "single_word": false,
433
+ "special": true
434
+ },
435
+ "256051": {
436
+ "content": "ewe_Latn",
437
+ "lstrip": false,
438
+ "normalized": false,
439
+ "rstrip": false,
440
+ "single_word": false,
441
+ "special": true
442
+ },
443
+ "256052": {
444
+ "content": "fao_Latn",
445
+ "lstrip": false,
446
+ "normalized": false,
447
+ "rstrip": false,
448
+ "single_word": false,
449
+ "special": true
450
+ },
451
+ "256053": {
452
+ "content": "pes_Arab",
453
+ "lstrip": false,
454
+ "normalized": false,
455
+ "rstrip": false,
456
+ "single_word": false,
457
+ "special": true
458
+ },
459
+ "256054": {
460
+ "content": "fij_Latn",
461
+ "lstrip": false,
462
+ "normalized": false,
463
+ "rstrip": false,
464
+ "single_word": false,
465
+ "special": true
466
+ },
467
+ "256055": {
468
+ "content": "fin_Latn",
469
+ "lstrip": false,
470
+ "normalized": false,
471
+ "rstrip": false,
472
+ "single_word": false,
473
+ "special": true
474
+ },
475
+ "256056": {
476
+ "content": "fon_Latn",
477
+ "lstrip": false,
478
+ "normalized": false,
479
+ "rstrip": false,
480
+ "single_word": false,
481
+ "special": true
482
+ },
483
+ "256057": {
484
+ "content": "fra_Latn",
485
+ "lstrip": false,
486
+ "normalized": false,
487
+ "rstrip": false,
488
+ "single_word": false,
489
+ "special": true
490
+ },
491
+ "256058": {
492
+ "content": "fur_Latn",
493
+ "lstrip": false,
494
+ "normalized": false,
495
+ "rstrip": false,
496
+ "single_word": false,
497
+ "special": true
498
+ },
499
+ "256059": {
500
+ "content": "fuv_Latn",
501
+ "lstrip": false,
502
+ "normalized": false,
503
+ "rstrip": false,
504
+ "single_word": false,
505
+ "special": true
506
+ },
507
+ "256060": {
508
+ "content": "gla_Latn",
509
+ "lstrip": false,
510
+ "normalized": false,
511
+ "rstrip": false,
512
+ "single_word": false,
513
+ "special": true
514
+ },
515
+ "256061": {
516
+ "content": "gle_Latn",
517
+ "lstrip": false,
518
+ "normalized": false,
519
+ "rstrip": false,
520
+ "single_word": false,
521
+ "special": true
522
+ },
523
+ "256062": {
524
+ "content": "glg_Latn",
525
+ "lstrip": false,
526
+ "normalized": false,
527
+ "rstrip": false,
528
+ "single_word": false,
529
+ "special": true
530
+ },
531
+ "256063": {
532
+ "content": "grn_Latn",
533
+ "lstrip": false,
534
+ "normalized": false,
535
+ "rstrip": false,
536
+ "single_word": false,
537
+ "special": true
538
+ },
539
+ "256064": {
540
+ "content": "guj_Gujr",
541
+ "lstrip": false,
542
+ "normalized": false,
543
+ "rstrip": false,
544
+ "single_word": false,
545
+ "special": true
546
+ },
547
+ "256065": {
548
+ "content": "hat_Latn",
549
+ "lstrip": false,
550
+ "normalized": false,
551
+ "rstrip": false,
552
+ "single_word": false,
553
+ "special": true
554
+ },
555
+ "256066": {
556
+ "content": "hau_Latn",
557
+ "lstrip": false,
558
+ "normalized": false,
559
+ "rstrip": false,
560
+ "single_word": false,
561
+ "special": true
562
+ },
563
+ "256067": {
564
+ "content": "heb_Hebr",
565
+ "lstrip": false,
566
+ "normalized": false,
567
+ "rstrip": false,
568
+ "single_word": false,
569
+ "special": true
570
+ },
571
+ "256068": {
572
+ "content": "hin_Deva",
573
+ "lstrip": false,
574
+ "normalized": false,
575
+ "rstrip": false,
576
+ "single_word": false,
577
+ "special": true
578
+ },
579
+ "256069": {
580
+ "content": "hne_Deva",
581
+ "lstrip": false,
582
+ "normalized": false,
583
+ "rstrip": false,
584
+ "single_word": false,
585
+ "special": true
586
+ },
587
+ "256070": {
588
+ "content": "hrv_Latn",
589
+ "lstrip": false,
590
+ "normalized": false,
591
+ "rstrip": false,
592
+ "single_word": false,
593
+ "special": true
594
+ },
595
+ "256071": {
596
+ "content": "hun_Latn",
597
+ "lstrip": false,
598
+ "normalized": false,
599
+ "rstrip": false,
600
+ "single_word": false,
601
+ "special": true
602
+ },
603
+ "256072": {
604
+ "content": "hye_Armn",
605
+ "lstrip": false,
606
+ "normalized": false,
607
+ "rstrip": false,
608
+ "single_word": false,
609
+ "special": true
610
+ },
611
+ "256073": {
612
+ "content": "ibo_Latn",
613
+ "lstrip": false,
614
+ "normalized": false,
615
+ "rstrip": false,
616
+ "single_word": false,
617
+ "special": true
618
+ },
619
+ "256074": {
620
+ "content": "ilo_Latn",
621
+ "lstrip": false,
622
+ "normalized": false,
623
+ "rstrip": false,
624
+ "single_word": false,
625
+ "special": true
626
+ },
627
+ "256075": {
628
+ "content": "ind_Latn",
629
+ "lstrip": false,
630
+ "normalized": false,
631
+ "rstrip": false,
632
+ "single_word": false,
633
+ "special": true
634
+ },
635
+ "256076": {
636
+ "content": "isl_Latn",
637
+ "lstrip": false,
638
+ "normalized": false,
639
+ "rstrip": false,
640
+ "single_word": false,
641
+ "special": true
642
+ },
643
+ "256077": {
644
+ "content": "ita_Latn",
645
+ "lstrip": false,
646
+ "normalized": false,
647
+ "rstrip": false,
648
+ "single_word": false,
649
+ "special": true
650
+ },
651
+ "256078": {
652
+ "content": "jav_Latn",
653
+ "lstrip": false,
654
+ "normalized": false,
655
+ "rstrip": false,
656
+ "single_word": false,
657
+ "special": true
658
+ },
659
+ "256079": {
660
+ "content": "jpn_Jpan",
661
+ "lstrip": false,
662
+ "normalized": false,
663
+ "rstrip": false,
664
+ "single_word": false,
665
+ "special": true
666
+ },
667
+ "256080": {
668
+ "content": "kab_Latn",
669
+ "lstrip": false,
670
+ "normalized": false,
671
+ "rstrip": false,
672
+ "single_word": false,
673
+ "special": true
674
+ },
675
+ "256081": {
676
+ "content": "kac_Latn",
677
+ "lstrip": false,
678
+ "normalized": false,
679
+ "rstrip": false,
680
+ "single_word": false,
681
+ "special": true
682
+ },
683
+ "256082": {
684
+ "content": "kam_Latn",
685
+ "lstrip": false,
686
+ "normalized": false,
687
+ "rstrip": false,
688
+ "single_word": false,
689
+ "special": true
690
+ },
691
+ "256083": {
692
+ "content": "kan_Knda",
693
+ "lstrip": false,
694
+ "normalized": false,
695
+ "rstrip": false,
696
+ "single_word": false,
697
+ "special": true
698
+ },
699
+ "256084": {
700
+ "content": "kas_Arab",
701
+ "lstrip": false,
702
+ "normalized": false,
703
+ "rstrip": false,
704
+ "single_word": false,
705
+ "special": true
706
+ },
707
+ "256085": {
708
+ "content": "kas_Deva",
709
+ "lstrip": false,
710
+ "normalized": false,
711
+ "rstrip": false,
712
+ "single_word": false,
713
+ "special": true
714
+ },
715
+ "256086": {
716
+ "content": "kat_Geor",
717
+ "lstrip": false,
718
+ "normalized": false,
719
+ "rstrip": false,
720
+ "single_word": false,
721
+ "special": true
722
+ },
723
+ "256087": {
724
+ "content": "knc_Arab",
725
+ "lstrip": false,
726
+ "normalized": false,
727
+ "rstrip": false,
728
+ "single_word": false,
729
+ "special": true
730
+ },
731
+ "256088": {
732
+ "content": "knc_Latn",
733
+ "lstrip": false,
734
+ "normalized": false,
735
+ "rstrip": false,
736
+ "single_word": false,
737
+ "special": true
738
+ },
739
+ "256089": {
740
+ "content": "kaz_Cyrl",
741
+ "lstrip": false,
742
+ "normalized": false,
743
+ "rstrip": false,
744
+ "single_word": false,
745
+ "special": true
746
+ },
747
+ "256090": {
748
+ "content": "kbp_Latn",
749
+ "lstrip": false,
750
+ "normalized": false,
751
+ "rstrip": false,
752
+ "single_word": false,
753
+ "special": true
754
+ },
755
+ "256091": {
756
+ "content": "kea_Latn",
757
+ "lstrip": false,
758
+ "normalized": false,
759
+ "rstrip": false,
760
+ "single_word": false,
761
+ "special": true
762
+ },
763
+ "256092": {
764
+ "content": "khm_Khmr",
765
+ "lstrip": false,
766
+ "normalized": false,
767
+ "rstrip": false,
768
+ "single_word": false,
769
+ "special": true
770
+ },
771
+ "256093": {
772
+ "content": "kik_Latn",
773
+ "lstrip": false,
774
+ "normalized": false,
775
+ "rstrip": false,
776
+ "single_word": false,
777
+ "special": true
778
+ },
779
+ "256094": {
780
+ "content": "kin_Latn",
781
+ "lstrip": false,
782
+ "normalized": false,
783
+ "rstrip": false,
784
+ "single_word": false,
785
+ "special": true
786
+ },
787
+ "256095": {
788
+ "content": "kir_Cyrl",
789
+ "lstrip": false,
790
+ "normalized": false,
791
+ "rstrip": false,
792
+ "single_word": false,
793
+ "special": true
794
+ },
795
+ "256096": {
796
+ "content": "kmb_Latn",
797
+ "lstrip": false,
798
+ "normalized": false,
799
+ "rstrip": false,
800
+ "single_word": false,
801
+ "special": true
802
+ },
803
+ "256097": {
804
+ "content": "kon_Latn",
805
+ "lstrip": false,
806
+ "normalized": false,
807
+ "rstrip": false,
808
+ "single_word": false,
809
+ "special": true
810
+ },
811
+ "256098": {
812
+ "content": "kor_Hang",
813
+ "lstrip": false,
814
+ "normalized": false,
815
+ "rstrip": false,
816
+ "single_word": false,
817
+ "special": true
818
+ },
819
+ "256099": {
820
+ "content": "kmr_Latn",
821
+ "lstrip": false,
822
+ "normalized": false,
823
+ "rstrip": false,
824
+ "single_word": false,
825
+ "special": true
826
+ },
827
+ "256100": {
828
+ "content": "lao_Laoo",
829
+ "lstrip": false,
830
+ "normalized": false,
831
+ "rstrip": false,
832
+ "single_word": false,
833
+ "special": true
834
+ },
835
+ "256101": {
836
+ "content": "lvs_Latn",
837
+ "lstrip": false,
838
+ "normalized": false,
839
+ "rstrip": false,
840
+ "single_word": false,
841
+ "special": true
842
+ },
843
+ "256102": {
844
+ "content": "lij_Latn",
845
+ "lstrip": false,
846
+ "normalized": false,
847
+ "rstrip": false,
848
+ "single_word": false,
849
+ "special": true
850
+ },
851
+ "256103": {
852
+ "content": "lim_Latn",
853
+ "lstrip": false,
854
+ "normalized": false,
855
+ "rstrip": false,
856
+ "single_word": false,
857
+ "special": true
858
+ },
859
+ "256104": {
860
+ "content": "lin_Latn",
861
+ "lstrip": false,
862
+ "normalized": false,
863
+ "rstrip": false,
864
+ "single_word": false,
865
+ "special": true
866
+ },
867
+ "256105": {
868
+ "content": "lit_Latn",
869
+ "lstrip": false,
870
+ "normalized": false,
871
+ "rstrip": false,
872
+ "single_word": false,
873
+ "special": true
874
+ },
875
+ "256106": {
876
+ "content": "lmo_Latn",
877
+ "lstrip": false,
878
+ "normalized": false,
879
+ "rstrip": false,
880
+ "single_word": false,
881
+ "special": true
882
+ },
883
+ "256107": {
884
+ "content": "ltg_Latn",
885
+ "lstrip": false,
886
+ "normalized": false,
887
+ "rstrip": false,
888
+ "single_word": false,
889
+ "special": true
890
+ },
891
+ "256108": {
892
+ "content": "ltz_Latn",
893
+ "lstrip": false,
894
+ "normalized": false,
895
+ "rstrip": false,
896
+ "single_word": false,
897
+ "special": true
898
+ },
899
+ "256109": {
900
+ "content": "lua_Latn",
901
+ "lstrip": false,
902
+ "normalized": false,
903
+ "rstrip": false,
904
+ "single_word": false,
905
+ "special": true
906
+ },
907
+ "256110": {
908
+ "content": "lug_Latn",
909
+ "lstrip": false,
910
+ "normalized": false,
911
+ "rstrip": false,
912
+ "single_word": false,
913
+ "special": true
914
+ },
915
+ "256111": {
916
+ "content": "luo_Latn",
917
+ "lstrip": false,
918
+ "normalized": false,
919
+ "rstrip": false,
920
+ "single_word": false,
921
+ "special": true
922
+ },
923
+ "256112": {
924
+ "content": "lus_Latn",
925
+ "lstrip": false,
926
+ "normalized": false,
927
+ "rstrip": false,
928
+ "single_word": false,
929
+ "special": true
930
+ },
931
+ "256113": {
932
+ "content": "mag_Deva",
933
+ "lstrip": false,
934
+ "normalized": false,
935
+ "rstrip": false,
936
+ "single_word": false,
937
+ "special": true
938
+ },
939
+ "256114": {
940
+ "content": "mai_Deva",
941
+ "lstrip": false,
942
+ "normalized": false,
943
+ "rstrip": false,
944
+ "single_word": false,
945
+ "special": true
946
+ },
947
+ "256115": {
948
+ "content": "mal_Mlym",
949
+ "lstrip": false,
950
+ "normalized": false,
951
+ "rstrip": false,
952
+ "single_word": false,
953
+ "special": true
954
+ },
955
+ "256116": {
956
+ "content": "mar_Deva",
957
+ "lstrip": false,
958
+ "normalized": false,
959
+ "rstrip": false,
960
+ "single_word": false,
961
+ "special": true
962
+ },
963
+ "256117": {
964
+ "content": "min_Latn",
965
+ "lstrip": false,
966
+ "normalized": false,
967
+ "rstrip": false,
968
+ "single_word": false,
969
+ "special": true
970
+ },
971
+ "256118": {
972
+ "content": "mkd_Cyrl",
973
+ "lstrip": false,
974
+ "normalized": false,
975
+ "rstrip": false,
976
+ "single_word": false,
977
+ "special": true
978
+ },
979
+ "256119": {
980
+ "content": "plt_Latn",
981
+ "lstrip": false,
982
+ "normalized": false,
983
+ "rstrip": false,
984
+ "single_word": false,
985
+ "special": true
986
+ },
987
+ "256120": {
988
+ "content": "mlt_Latn",
989
+ "lstrip": false,
990
+ "normalized": false,
991
+ "rstrip": false,
992
+ "single_word": false,
993
+ "special": true
994
+ },
995
+ "256121": {
996
+ "content": "mni_Beng",
997
+ "lstrip": false,
998
+ "normalized": false,
999
+ "rstrip": false,
1000
+ "single_word": false,
1001
+ "special": true
1002
+ },
1003
+ "256122": {
1004
+ "content": "khk_Cyrl",
1005
+ "lstrip": false,
1006
+ "normalized": false,
1007
+ "rstrip": false,
1008
+ "single_word": false,
1009
+ "special": true
1010
+ },
1011
+ "256123": {
1012
+ "content": "mos_Latn",
1013
+ "lstrip": false,
1014
+ "normalized": false,
1015
+ "rstrip": false,
1016
+ "single_word": false,
1017
+ "special": true
1018
+ },
1019
+ "256124": {
1020
+ "content": "mri_Latn",
1021
+ "lstrip": false,
1022
+ "normalized": false,
1023
+ "rstrip": false,
1024
+ "single_word": false,
1025
+ "special": true
1026
+ },
1027
+ "256125": {
1028
+ "content": "zsm_Latn",
1029
+ "lstrip": false,
1030
+ "normalized": false,
1031
+ "rstrip": false,
1032
+ "single_word": false,
1033
+ "special": true
1034
+ },
1035
+ "256126": {
1036
+ "content": "mya_Mymr",
1037
+ "lstrip": false,
1038
+ "normalized": false,
1039
+ "rstrip": false,
1040
+ "single_word": false,
1041
+ "special": true
1042
+ },
1043
+ "256127": {
1044
+ "content": "nld_Latn",
1045
+ "lstrip": false,
1046
+ "normalized": false,
1047
+ "rstrip": false,
1048
+ "single_word": false,
1049
+ "special": true
1050
+ },
1051
+ "256128": {
1052
+ "content": "nno_Latn",
1053
+ "lstrip": false,
1054
+ "normalized": false,
1055
+ "rstrip": false,
1056
+ "single_word": false,
1057
+ "special": true
1058
+ },
1059
+ "256129": {
1060
+ "content": "nob_Latn",
1061
+ "lstrip": false,
1062
+ "normalized": false,
1063
+ "rstrip": false,
1064
+ "single_word": false,
1065
+ "special": true
1066
+ },
1067
+ "256130": {
1068
+ "content": "npi_Deva",
1069
+ "lstrip": false,
1070
+ "normalized": false,
1071
+ "rstrip": false,
1072
+ "single_word": false,
1073
+ "special": true
1074
+ },
1075
+ "256131": {
1076
+ "content": "nso_Latn",
1077
+ "lstrip": false,
1078
+ "normalized": false,
1079
+ "rstrip": false,
1080
+ "single_word": false,
1081
+ "special": true
1082
+ },
1083
+ "256132": {
1084
+ "content": "nus_Latn",
1085
+ "lstrip": false,
1086
+ "normalized": false,
1087
+ "rstrip": false,
1088
+ "single_word": false,
1089
+ "special": true
1090
+ },
1091
+ "256133": {
1092
+ "content": "nya_Latn",
1093
+ "lstrip": false,
1094
+ "normalized": false,
1095
+ "rstrip": false,
1096
+ "single_word": false,
1097
+ "special": true
1098
+ },
1099
+ "256134": {
1100
+ "content": "oci_Latn",
1101
+ "lstrip": false,
1102
+ "normalized": false,
1103
+ "rstrip": false,
1104
+ "single_word": false,
1105
+ "special": true
1106
+ },
1107
+ "256135": {
1108
+ "content": "gaz_Latn",
1109
+ "lstrip": false,
1110
+ "normalized": false,
1111
+ "rstrip": false,
1112
+ "single_word": false,
1113
+ "special": true
1114
+ },
1115
+ "256136": {
1116
+ "content": "ory_Orya",
1117
+ "lstrip": false,
1118
+ "normalized": false,
1119
+ "rstrip": false,
1120
+ "single_word": false,
1121
+ "special": true
1122
+ },
1123
+ "256137": {
1124
+ "content": "pag_Latn",
1125
+ "lstrip": false,
1126
+ "normalized": false,
1127
+ "rstrip": false,
1128
+ "single_word": false,
1129
+ "special": true
1130
+ },
1131
+ "256138": {
1132
+ "content": "pan_Guru",
1133
+ "lstrip": false,
1134
+ "normalized": false,
1135
+ "rstrip": false,
1136
+ "single_word": false,
1137
+ "special": true
1138
+ },
1139
+ "256139": {
1140
+ "content": "pap_Latn",
1141
+ "lstrip": false,
1142
+ "normalized": false,
1143
+ "rstrip": false,
1144
+ "single_word": false,
1145
+ "special": true
1146
+ },
1147
+ "256140": {
1148
+ "content": "pol_Latn",
1149
+ "lstrip": false,
1150
+ "normalized": false,
1151
+ "rstrip": false,
1152
+ "single_word": false,
1153
+ "special": true
1154
+ },
1155
+ "256141": {
1156
+ "content": "por_Latn",
1157
+ "lstrip": false,
1158
+ "normalized": false,
1159
+ "rstrip": false,
1160
+ "single_word": false,
1161
+ "special": true
1162
+ },
1163
+ "256142": {
1164
+ "content": "prs_Arab",
1165
+ "lstrip": false,
1166
+ "normalized": false,
1167
+ "rstrip": false,
1168
+ "single_word": false,
1169
+ "special": true
1170
+ },
1171
+ "256143": {
1172
+ "content": "pbt_Arab",
1173
+ "lstrip": false,
1174
+ "normalized": false,
1175
+ "rstrip": false,
1176
+ "single_word": false,
1177
+ "special": true
1178
+ },
1179
+ "256144": {
1180
+ "content": "quy_Latn",
1181
+ "lstrip": false,
1182
+ "normalized": false,
1183
+ "rstrip": false,
1184
+ "single_word": false,
1185
+ "special": true
1186
+ },
1187
+ "256145": {
1188
+ "content": "ron_Latn",
1189
+ "lstrip": false,
1190
+ "normalized": false,
1191
+ "rstrip": false,
1192
+ "single_word": false,
1193
+ "special": true
1194
+ },
1195
+ "256146": {
1196
+ "content": "run_Latn",
1197
+ "lstrip": false,
1198
+ "normalized": false,
1199
+ "rstrip": false,
1200
+ "single_word": false,
1201
+ "special": true
1202
+ },
1203
+ "256147": {
1204
+ "content": "rus_Cyrl",
1205
+ "lstrip": false,
1206
+ "normalized": false,
1207
+ "rstrip": false,
1208
+ "single_word": false,
1209
+ "special": true
1210
+ },
1211
+ "256148": {
1212
+ "content": "sag_Latn",
1213
+ "lstrip": false,
1214
+ "normalized": false,
1215
+ "rstrip": false,
1216
+ "single_word": false,
1217
+ "special": true
1218
+ },
1219
+ "256149": {
1220
+ "content": "san_Deva",
1221
+ "lstrip": false,
1222
+ "normalized": false,
1223
+ "rstrip": false,
1224
+ "single_word": false,
1225
+ "special": true
1226
+ },
1227
+ "256150": {
1228
+ "content": "sat_Beng",
1229
+ "lstrip": false,
1230
+ "normalized": false,
1231
+ "rstrip": false,
1232
+ "single_word": false,
1233
+ "special": true
1234
+ },
1235
+ "256151": {
1236
+ "content": "scn_Latn",
1237
+ "lstrip": false,
1238
+ "normalized": false,
1239
+ "rstrip": false,
1240
+ "single_word": false,
1241
+ "special": true
1242
+ },
1243
+ "256152": {
1244
+ "content": "shn_Mymr",
1245
+ "lstrip": false,
1246
+ "normalized": false,
1247
+ "rstrip": false,
1248
+ "single_word": false,
1249
+ "special": true
1250
+ },
1251
+ "256153": {
1252
+ "content": "sin_Sinh",
1253
+ "lstrip": false,
1254
+ "normalized": false,
1255
+ "rstrip": false,
1256
+ "single_word": false,
1257
+ "special": true
1258
+ },
1259
+ "256154": {
1260
+ "content": "slk_Latn",
1261
+ "lstrip": false,
1262
+ "normalized": false,
1263
+ "rstrip": false,
1264
+ "single_word": false,
1265
+ "special": true
1266
+ },
1267
+ "256155": {
1268
+ "content": "slv_Latn",
1269
+ "lstrip": false,
1270
+ "normalized": false,
1271
+ "rstrip": false,
1272
+ "single_word": false,
1273
+ "special": true
1274
+ },
1275
+ "256156": {
1276
+ "content": "smo_Latn",
1277
+ "lstrip": false,
1278
+ "normalized": false,
1279
+ "rstrip": false,
1280
+ "single_word": false,
1281
+ "special": true
1282
+ },
1283
+ "256157": {
1284
+ "content": "sna_Latn",
1285
+ "lstrip": false,
1286
+ "normalized": false,
1287
+ "rstrip": false,
1288
+ "single_word": false,
1289
+ "special": true
1290
+ },
1291
+ "256158": {
1292
+ "content": "snd_Arab",
1293
+ "lstrip": false,
1294
+ "normalized": false,
1295
+ "rstrip": false,
1296
+ "single_word": false,
1297
+ "special": true
1298
+ },
1299
+ "256159": {
1300
+ "content": "som_Latn",
1301
+ "lstrip": false,
1302
+ "normalized": false,
1303
+ "rstrip": false,
1304
+ "single_word": false,
1305
+ "special": true
1306
+ },
1307
+ "256160": {
1308
+ "content": "sot_Latn",
1309
+ "lstrip": false,
1310
+ "normalized": false,
1311
+ "rstrip": false,
1312
+ "single_word": false,
1313
+ "special": true
1314
+ },
1315
+ "256161": {
1316
+ "content": "spa_Latn",
1317
+ "lstrip": false,
1318
+ "normalized": false,
1319
+ "rstrip": false,
1320
+ "single_word": false,
1321
+ "special": true
1322
+ },
1323
+ "256162": {
1324
+ "content": "als_Latn",
1325
+ "lstrip": false,
1326
+ "normalized": false,
1327
+ "rstrip": false,
1328
+ "single_word": false,
1329
+ "special": true
1330
+ },
1331
+ "256163": {
1332
+ "content": "srd_Latn",
1333
+ "lstrip": false,
1334
+ "normalized": false,
1335
+ "rstrip": false,
1336
+ "single_word": false,
1337
+ "special": true
1338
+ },
1339
+ "256164": {
1340
+ "content": "srp_Cyrl",
1341
+ "lstrip": false,
1342
+ "normalized": false,
1343
+ "rstrip": false,
1344
+ "single_word": false,
1345
+ "special": true
1346
+ },
1347
+ "256165": {
1348
+ "content": "ssw_Latn",
1349
+ "lstrip": false,
1350
+ "normalized": false,
1351
+ "rstrip": false,
1352
+ "single_word": false,
1353
+ "special": true
1354
+ },
1355
+ "256166": {
1356
+ "content": "sun_Latn",
1357
+ "lstrip": false,
1358
+ "normalized": false,
1359
+ "rstrip": false,
1360
+ "single_word": false,
1361
+ "special": true
1362
+ },
1363
+ "256167": {
1364
+ "content": "swe_Latn",
1365
+ "lstrip": false,
1366
+ "normalized": false,
1367
+ "rstrip": false,
1368
+ "single_word": false,
1369
+ "special": true
1370
+ },
1371
+ "256168": {
1372
+ "content": "swh_Latn",
1373
+ "lstrip": false,
1374
+ "normalized": false,
1375
+ "rstrip": false,
1376
+ "single_word": false,
1377
+ "special": true
1378
+ },
1379
+ "256169": {
1380
+ "content": "szl_Latn",
1381
+ "lstrip": false,
1382
+ "normalized": false,
1383
+ "rstrip": false,
1384
+ "single_word": false,
1385
+ "special": true
1386
+ },
1387
+ "256170": {
1388
+ "content": "tam_Taml",
1389
+ "lstrip": false,
1390
+ "normalized": false,
1391
+ "rstrip": false,
1392
+ "single_word": false,
1393
+ "special": true
1394
+ },
1395
+ "256171": {
1396
+ "content": "tat_Cyrl",
1397
+ "lstrip": false,
1398
+ "normalized": false,
1399
+ "rstrip": false,
1400
+ "single_word": false,
1401
+ "special": true
1402
+ },
1403
+ "256172": {
1404
+ "content": "tel_Telu",
1405
+ "lstrip": false,
1406
+ "normalized": false,
1407
+ "rstrip": false,
1408
+ "single_word": false,
1409
+ "special": true
1410
+ },
1411
+ "256173": {
1412
+ "content": "tgk_Cyrl",
1413
+ "lstrip": false,
1414
+ "normalized": false,
1415
+ "rstrip": false,
1416
+ "single_word": false,
1417
+ "special": true
1418
+ },
1419
+ "256174": {
1420
+ "content": "tgl_Latn",
1421
+ "lstrip": false,
1422
+ "normalized": false,
1423
+ "rstrip": false,
1424
+ "single_word": false,
1425
+ "special": true
1426
+ },
1427
+ "256175": {
1428
+ "content": "tha_Thai",
1429
+ "lstrip": false,
1430
+ "normalized": false,
1431
+ "rstrip": false,
1432
+ "single_word": false,
1433
+ "special": true
1434
+ },
1435
+ "256176": {
1436
+ "content": "tir_Ethi",
1437
+ "lstrip": false,
1438
+ "normalized": false,
1439
+ "rstrip": false,
1440
+ "single_word": false,
1441
+ "special": true
1442
+ },
1443
+ "256177": {
1444
+ "content": "taq_Latn",
1445
+ "lstrip": false,
1446
+ "normalized": false,
1447
+ "rstrip": false,
1448
+ "single_word": false,
1449
+ "special": true
1450
+ },
1451
+ "256178": {
1452
+ "content": "taq_Tfng",
1453
+ "lstrip": false,
1454
+ "normalized": false,
1455
+ "rstrip": false,
1456
+ "single_word": false,
1457
+ "special": true
1458
+ },
1459
+ "256179": {
1460
+ "content": "tpi_Latn",
1461
+ "lstrip": false,
1462
+ "normalized": false,
1463
+ "rstrip": false,
1464
+ "single_word": false,
1465
+ "special": true
1466
+ },
1467
+ "256180": {
1468
+ "content": "tsn_Latn",
1469
+ "lstrip": false,
1470
+ "normalized": false,
1471
+ "rstrip": false,
1472
+ "single_word": false,
1473
+ "special": true
1474
+ },
1475
+ "256181": {
1476
+ "content": "tso_Latn",
1477
+ "lstrip": false,
1478
+ "normalized": false,
1479
+ "rstrip": false,
1480
+ "single_word": false,
1481
+ "special": true
1482
+ },
1483
+ "256182": {
1484
+ "content": "tuk_Latn",
1485
+ "lstrip": false,
1486
+ "normalized": false,
1487
+ "rstrip": false,
1488
+ "single_word": false,
1489
+ "special": true
1490
+ },
1491
+ "256183": {
1492
+ "content": "tum_Latn",
1493
+ "lstrip": false,
1494
+ "normalized": false,
1495
+ "rstrip": false,
1496
+ "single_word": false,
1497
+ "special": true
1498
+ },
1499
+ "256184": {
1500
+ "content": "tur_Latn",
1501
+ "lstrip": false,
1502
+ "normalized": false,
1503
+ "rstrip": false,
1504
+ "single_word": false,
1505
+ "special": true
1506
+ },
1507
+ "256185": {
1508
+ "content": "twi_Latn",
1509
+ "lstrip": false,
1510
+ "normalized": false,
1511
+ "rstrip": false,
1512
+ "single_word": false,
1513
+ "special": true
1514
+ },
1515
+ "256186": {
1516
+ "content": "tzm_Tfng",
1517
+ "lstrip": false,
1518
+ "normalized": false,
1519
+ "rstrip": false,
1520
+ "single_word": false,
1521
+ "special": true
1522
+ },
1523
+ "256187": {
1524
+ "content": "uig_Arab",
1525
+ "lstrip": false,
1526
+ "normalized": false,
1527
+ "rstrip": false,
1528
+ "single_word": false,
1529
+ "special": true
1530
+ },
1531
+ "256188": {
1532
+ "content": "ukr_Cyrl",
1533
+ "lstrip": false,
1534
+ "normalized": false,
1535
+ "rstrip": false,
1536
+ "single_word": false,
1537
+ "special": true
1538
+ },
1539
+ "256189": {
1540
+ "content": "umb_Latn",
1541
+ "lstrip": false,
1542
+ "normalized": false,
1543
+ "rstrip": false,
1544
+ "single_word": false,
1545
+ "special": true
1546
+ },
1547
+ "256190": {
1548
+ "content": "urd_Arab",
1549
+ "lstrip": false,
1550
+ "normalized": false,
1551
+ "rstrip": false,
1552
+ "single_word": false,
1553
+ "special": true
1554
+ },
1555
+ "256191": {
1556
+ "content": "uzn_Latn",
1557
+ "lstrip": false,
1558
+ "normalized": false,
1559
+ "rstrip": false,
1560
+ "single_word": false,
1561
+ "special": true
1562
+ },
1563
+ "256192": {
1564
+ "content": "vec_Latn",
1565
+ "lstrip": false,
1566
+ "normalized": false,
1567
+ "rstrip": false,
1568
+ "single_word": false,
1569
+ "special": true
1570
+ },
1571
+ "256193": {
1572
+ "content": "vie_Latn",
1573
+ "lstrip": false,
1574
+ "normalized": false,
1575
+ "rstrip": false,
1576
+ "single_word": false,
1577
+ "special": true
1578
+ },
1579
+ "256194": {
1580
+ "content": "war_Latn",
1581
+ "lstrip": false,
1582
+ "normalized": false,
1583
+ "rstrip": false,
1584
+ "single_word": false,
1585
+ "special": true
1586
+ },
1587
+ "256195": {
1588
+ "content": "wol_Latn",
1589
+ "lstrip": false,
1590
+ "normalized": false,
1591
+ "rstrip": false,
1592
+ "single_word": false,
1593
+ "special": true
1594
+ },
1595
+ "256196": {
1596
+ "content": "xho_Latn",
1597
+ "lstrip": false,
1598
+ "normalized": false,
1599
+ "rstrip": false,
1600
+ "single_word": false,
1601
+ "special": true
1602
+ },
1603
+ "256197": {
1604
+ "content": "ydd_Hebr",
1605
+ "lstrip": false,
1606
+ "normalized": false,
1607
+ "rstrip": false,
1608
+ "single_word": false,
1609
+ "special": true
1610
+ },
1611
+ "256198": {
1612
+ "content": "yor_Latn",
1613
+ "lstrip": false,
1614
+ "normalized": false,
1615
+ "rstrip": false,
1616
+ "single_word": false,
1617
+ "special": true
1618
+ },
1619
+ "256199": {
1620
+ "content": "yue_Hant",
1621
+ "lstrip": false,
1622
+ "normalized": false,
1623
+ "rstrip": false,
1624
+ "single_word": false,
1625
+ "special": true
1626
+ },
1627
+ "256200": {
1628
+ "content": "zho_Hans",
1629
+ "lstrip": false,
1630
+ "normalized": false,
1631
+ "rstrip": false,
1632
+ "single_word": false,
1633
+ "special": true
1634
+ },
1635
+ "256201": {
1636
+ "content": "zho_Hant",
1637
+ "lstrip": false,
1638
+ "normalized": false,
1639
+ "rstrip": false,
1640
+ "single_word": false,
1641
+ "special": true
1642
+ },
1643
+ "256202": {
1644
+ "content": "zul_Latn",
1645
+ "lstrip": false,
1646
+ "normalized": false,
1647
+ "rstrip": false,
1648
+ "single_word": false,
1649
+ "special": true
1650
+ },
1651
+ "256203": {
1652
+ "content": "<mask>",
1653
+ "lstrip": true,
1654
+ "normalized": true,
1655
+ "rstrip": false,
1656
+ "single_word": false,
1657
+ "special": true
1658
+ }
1659
+ },
1660
+ "additional_special_tokens": [
1661
+ "ace_Arab",
1662
+ "ace_Latn",
1663
+ "acm_Arab",
1664
+ "acq_Arab",
1665
+ "aeb_Arab",
1666
+ "afr_Latn",
1667
+ "ajp_Arab",
1668
+ "aka_Latn",
1669
+ "amh_Ethi",
1670
+ "apc_Arab",
1671
+ "arb_Arab",
1672
+ "ars_Arab",
1673
+ "ary_Arab",
1674
+ "arz_Arab",
1675
+ "asm_Beng",
1676
+ "ast_Latn",
1677
+ "awa_Deva",
1678
+ "ayr_Latn",
1679
+ "azb_Arab",
1680
+ "azj_Latn",
1681
+ "bak_Cyrl",
1682
+ "bam_Latn",
1683
+ "ban_Latn",
1684
+ "bel_Cyrl",
1685
+ "bem_Latn",
1686
+ "ben_Beng",
1687
+ "bho_Deva",
1688
+ "bjn_Arab",
1689
+ "bjn_Latn",
1690
+ "bod_Tibt",
1691
+ "bos_Latn",
1692
+ "bug_Latn",
1693
+ "bul_Cyrl",
1694
+ "cat_Latn",
1695
+ "ceb_Latn",
1696
+ "ces_Latn",
1697
+ "cjk_Latn",
1698
+ "ckb_Arab",
1699
+ "crh_Latn",
1700
+ "cym_Latn",
1701
+ "dan_Latn",
1702
+ "deu_Latn",
1703
+ "dik_Latn",
1704
+ "dyu_Latn",
1705
+ "dzo_Tibt",
1706
+ "ell_Grek",
1707
+ "eng_Latn",
1708
+ "epo_Latn",
1709
+ "est_Latn",
1710
+ "eus_Latn",
1711
+ "ewe_Latn",
1712
+ "fao_Latn",
1713
+ "pes_Arab",
1714
+ "fij_Latn",
1715
+ "fin_Latn",
1716
+ "fon_Latn",
1717
+ "fra_Latn",
1718
+ "fur_Latn",
1719
+ "fuv_Latn",
1720
+ "gla_Latn",
1721
+ "gle_Latn",
1722
+ "glg_Latn",
1723
+ "grn_Latn",
1724
+ "guj_Gujr",
1725
+ "hat_Latn",
1726
+ "hau_Latn",
1727
+ "heb_Hebr",
1728
+ "hin_Deva",
1729
+ "hne_Deva",
1730
+ "hrv_Latn",
1731
+ "hun_Latn",
1732
+ "hye_Armn",
1733
+ "ibo_Latn",
1734
+ "ilo_Latn",
1735
+ "ind_Latn",
1736
+ "isl_Latn",
1737
+ "ita_Latn",
1738
+ "jav_Latn",
1739
+ "jpn_Jpan",
1740
+ "kab_Latn",
1741
+ "kac_Latn",
1742
+ "kam_Latn",
1743
+ "kan_Knda",
1744
+ "kas_Arab",
1745
+ "kas_Deva",
1746
+ "kat_Geor",
1747
+ "knc_Arab",
1748
+ "knc_Latn",
1749
+ "kaz_Cyrl",
1750
+ "kbp_Latn",
1751
+ "kea_Latn",
1752
+ "khm_Khmr",
1753
+ "kik_Latn",
1754
+ "kin_Latn",
1755
+ "kir_Cyrl",
1756
+ "kmb_Latn",
1757
+ "kon_Latn",
1758
+ "kor_Hang",
1759
+ "kmr_Latn",
1760
+ "lao_Laoo",
1761
+ "lvs_Latn",
1762
+ "lij_Latn",
1763
+ "lim_Latn",
1764
+ "lin_Latn",
1765
+ "lit_Latn",
1766
+ "lmo_Latn",
1767
+ "ltg_Latn",
1768
+ "ltz_Latn",
1769
+ "lua_Latn",
1770
+ "lug_Latn",
1771
+ "luo_Latn",
1772
+ "lus_Latn",
1773
+ "mag_Deva",
1774
+ "mai_Deva",
1775
+ "mal_Mlym",
1776
+ "mar_Deva",
1777
+ "min_Latn",
1778
+ "mkd_Cyrl",
1779
+ "plt_Latn",
1780
+ "mlt_Latn",
1781
+ "mni_Beng",
1782
+ "khk_Cyrl",
1783
+ "mos_Latn",
1784
+ "mri_Latn",
1785
+ "zsm_Latn",
1786
+ "mya_Mymr",
1787
+ "nld_Latn",
1788
+ "nno_Latn",
1789
+ "nob_Latn",
1790
+ "npi_Deva",
1791
+ "nso_Latn",
1792
+ "nus_Latn",
1793
+ "nya_Latn",
1794
+ "oci_Latn",
1795
+ "gaz_Latn",
1796
+ "ory_Orya",
1797
+ "pag_Latn",
1798
+ "pan_Guru",
1799
+ "pap_Latn",
1800
+ "pol_Latn",
1801
+ "por_Latn",
1802
+ "prs_Arab",
1803
+ "pbt_Arab",
1804
+ "quy_Latn",
1805
+ "ron_Latn",
1806
+ "run_Latn",
1807
+ "rus_Cyrl",
1808
+ "sag_Latn",
1809
+ "san_Deva",
1810
+ "sat_Beng",
1811
+ "scn_Latn",
1812
+ "shn_Mymr",
1813
+ "sin_Sinh",
1814
+ "slk_Latn",
1815
+ "slv_Latn",
1816
+ "smo_Latn",
1817
+ "sna_Latn",
1818
+ "snd_Arab",
1819
+ "som_Latn",
1820
+ "sot_Latn",
1821
+ "spa_Latn",
1822
+ "als_Latn",
1823
+ "srd_Latn",
1824
+ "srp_Cyrl",
1825
+ "ssw_Latn",
1826
+ "sun_Latn",
1827
+ "swe_Latn",
1828
+ "swh_Latn",
1829
+ "szl_Latn",
1830
+ "tam_Taml",
1831
+ "tat_Cyrl",
1832
+ "tel_Telu",
1833
+ "tgk_Cyrl",
1834
+ "tgl_Latn",
1835
+ "tha_Thai",
1836
+ "tir_Ethi",
1837
+ "taq_Latn",
1838
+ "taq_Tfng",
1839
+ "tpi_Latn",
1840
+ "tsn_Latn",
1841
+ "tso_Latn",
1842
+ "tuk_Latn",
1843
+ "tum_Latn",
1844
+ "tur_Latn",
1845
+ "twi_Latn",
1846
+ "tzm_Tfng",
1847
+ "uig_Arab",
1848
+ "ukr_Cyrl",
1849
+ "umb_Latn",
1850
+ "urd_Arab",
1851
+ "uzn_Latn",
1852
+ "vec_Latn",
1853
+ "vie_Latn",
1854
+ "war_Latn",
1855
+ "wol_Latn",
1856
+ "xho_Latn",
1857
+ "ydd_Hebr",
1858
+ "yor_Latn",
1859
+ "yue_Hant",
1860
+ "zho_Hans",
1861
+ "zho_Hant",
1862
+ "zul_Latn"
1863
+ ],
1864
+ "bos_token": "<s>",
1865
+ "clean_up_tokenization_spaces": false,
1866
+ "cls_token": "<s>",
1867
+ "eos_token": "</s>",
1868
+ "legacy_behaviour": false,
1869
+ "mask_token": "<mask>",
1870
+ "model_max_length": 1024,
1871
+ "pad_token": "<pad>",
1872
+ "sep_token": "</s>",
1873
+ "sp_model_kwargs": {},
1874
+ "src_lang": "eng_Latn",
1875
+ "tgt_lang": "mni_Beng",
1876
+ "tokenizer_class": "NllbTokenizer",
1877
+ "unk_token": "<unk>"
1878
+ }
nllb-mni-lsftl/checkpoint-26112/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
nltk_data/corpora/omw-1.4.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3b941e664852f3297b6040236626065796a2aaf7d7f9eec8779a3beaa1096c2d
3
+ size 26634772