 Matthias Reso
|
2e768b1d1d
Added unit test coverage for amp
|
hai 1 ano |
|
Matthias Reso
|
33925f71e6
Add missing amp context if use_fp16 is enabled
|
hai 1 ano |
 Hamid Shojanazeri
|
35b394e49f
adding profiler and flop_counter
|
hai 1 ano |
|
Hamid Shojanazeri
|
d56d5c469d
adding flop counter
|
hai 1 ano |
|
Hamid Shojanazeri
|
1ba9177645
minor updates (#228)
|
hai 1 ano |
 amitsangani
|
0190444849
minor updates
|
hai 1 ano |
|
Hamid Shojanazeri
|
e25bd3be25
Llama 2 Connect 2023 Colab Notebook (#227)
|
hai 1 ano |
|
amitsangani
|
b5267c515d
address comments
|
hai 1 ano |
|
Hamid Shojanazeri
|
3c620b2b29
clean up
|
hai 1 ano |
|
Hamid Shojanazeri
|
cd3a72ec35
fix format and missing packages
|
hai 1 ano |
|
amitsangani
|
636fd00d16
Llama 2 Connect 2023 Colab Notebook
|
hai 1 ano |
 Beto
|
7e7c8d2514
fixing typo on x label
|
hai 1 ano |
|
Beto
|
5a011246e1
Modularizing the graphs further. Plotting train and validation separetly due to difference in number of steps.
|
hai 1 ano |
|
Beto
|
9938afd088
condensed functionality in a single function for plotting both epoch and step metrics
|
hai 1 ano |
|
Matthias Reso
|
a0cd3c7c77
Added dependency to qsnb
|
hai 1 ano |
|
Matthias Reso
|
8077a08d78
Updated quickstart notebook to use llama_recipes package
|
hai 1 ano |
|
Beto
|
92be45b0fe
Adding matplotlib to requirements. Removing import from train_utils
|
hai 1 ano |
 Anshika
|
b8d327f7c3
Update grammar_dataset_process.ipynb
|
hai 1 ano |
|
Anshika
|
fecfece7e3
Update grammar_dataset_process.ipynb
|
hai 1 ano |
|
Beto
|
80a6ac41b5
Adding the rank to the metrics filename to differentiate the run from each GPU
|
hai 1 ano |
|
Beto
|
2f7256918e
Adding support to save fine tuning metrics to json. Adding script to plot saved metrics.
|
hai 1 ano |
 Geeta Chauhan
|
18e0198626
Make tests run on cpu instance (#212)
|
hai 1 ano |
|
Matthias Reso
|
584bc93609
Make tests run on cpu instance
|
hai 1 ano |
|
Matthias Reso
|
eafea7b366
Invalidate labels in dialog dataset to disable loss
|
hai 1 ano |
|
Matthias Reso
|
cc8cc0d3c3
fix grammar dataset
|
hai 1 ano |
|
Matthias Reso
|
2e4bd2a665
Resize vocab size to fix idx error
|
hai 1 ano |
|
Matthias Reso
|
10f9367e56
fix missing labels in datasets
|
hai 1 ano |
|
Matthias Reso
|
f2d02a9362
Add unit test for dis sampler
|
hai 1 ano |
|
Matthias Reso
|
be63d9ec39
Remove padding in alpaca ds; remove concat in grammar
|
hai 1 ano |
|
Matthias Reso
|
ddf58d205d
Added dist length based batch sampler
|
hai 1 ano |
