bertin-project/bertin-roberta-base-spanish at 73f6af597700170c72efb0ccb91daaa4b0ba1c05

bertin-roberta-base-spanish

8 contributors

History: 41 commits

versae's picture

Model at 182k steps, mlm acc 0.6494

73f6af5 over 2 years ago

configs
Changed and added vocab and tokenizer over 2 years ago
mc4
Fixes to mc4 fork over 2 years ago
.gitattributes

736 Bytes

Update .gitattributes over 2 years ago
.gitignore

1.84 kB

Initial test with BETO's corpus over 2 years ago
README.md

1.84 kB

Fixed widget example over 2 years ago
config.json

618 Bytes

Fix config for checkpoint over 2 years ago
config.py

247 Bytes

Scripts for perplexity sampling and fixes over 2 years ago
convert.py

469 Bytes

Adding missing import over 2 years ago
flax_model.msgpack

499 MB
LFS

Model at 182k steps, mlm acc 0.6494 over 2 years ago
get_embeddings_and_perplexity.py

1.53 kB

Add script to generate dataset of embeddings and perplexities. Add script to generate t-SNE plot for embedding and perplexity visualization. over 2 years ago
merges.txt

505 kB

Changed and added vocab and tokenizer over 2 years ago
perplexity.py

751 Bytes

Adding checkpointing, wandb, and new mlm script over 2 years ago
pytorch_model.bin
Detected Pickle imports (4)
- "torch.FloatStorage",
- "torch.LongStorage",
- "torch._utils._rebuild_tensor_v2",
- "collections.OrderedDict"
What is a pickle import?
499 MB
LFS

Model at 182k steps, mlm acc 0.6494 over 2 years ago
run.sh

883 Bytes

Adding base config and organizing configs over 2 years ago
run_mlm_flax.py

30 kB

Adding sampling to mc4 over 2 years ago
run_mlm_flax_stream.py

28.4 kB

Fixes to mc4 fork over 2 years ago
run_stream.sh

924 Bytes

Adding base config and organizing configs over 2 years ago
special_tokens_map.json

239 Bytes

Changed and added vocab and tokenizer over 2 years ago
tokenizer.json

1.45 MB

Changed and added vocab and tokenizer over 2 years ago
tokenizer_config.json

292 Bytes

Changed and added vocab and tokenizer over 2 years ago
tokens.py

649 Bytes

Scripts for perplexity sampling and fixes over 2 years ago
tokens.py.orig

899 Bytes

Adjust batch size for extrating tokens over 2 years ago
tsne_plot.py

3.02 kB

Remove unused imports over 2 years ago
vocab.json

846 kB

Changed and added vocab and tokenizer over 2 years ago