Duplicated from cfahlgren1/enron-autocomplete

paperweight
/

tinynext

Text Generation

fill-in-the-middle

onnxruntime-web

Eval Results (legacy)

Model card Files Files and versions

269 MB

Ctrl+K

Ctrl+K

1 contributor

History: 14 commits

cfahlgren1's picture

cfahlgren1 HF Staff

link the new paperweight home and default train.py to it

6f8d8d2 verified about 13 hours ago

experiments
remove the 3b-token experiment folder, now the same as v2 about 13 hours ago
metrics
make the 3b-token run v2 and add the size comparison about 16 hours ago
web
make the 3b-token run v2 and add the size comparison about 16 hours ago
compress.py

5.61 kB
v2: 13 MB browser build (int8 download, float compute), export/compress scripts, onnx evals about 22 hours ago
config.json

312 Bytes
tinynext-v2: 12.5M params, 28 min on GPU about 22 hours ago
eval.py

22.6 kB
make the 3b-token run v2 and add the size comparison about 16 hours ago
eval_set.json

275 kB
v2: training script, eval harness, eval set and baseline metrics 1 day ago
export.py

2.09 kB
export the compressed browser format from train.py instead of int8 activations about 21 hours ago
model.onnx

60.8 MB
xet

make the 3b-token run v2 and add the size comparison about 16 hours ago
model.safetensors

50.2 MB
xet

make the 3b-token run v2 and add the size comparison about 16 hours ago
model_int8.onnx

15.6 MB
xet

make the 3b-token run v2 and add the size comparison about 16 hours ago
report.json

47.8 kB
make the 3b-token run v2 and add the size comparison about 16 hours ago
tokenizer.json

551 kB
make the 3b-token run v2 and add the size comparison about 16 hours ago
train.py

34.2 kB
link the new paperweight home and default train.py to it about 13 hours ago