 Matthias Reso
|
8b0a233c1a
Use new chat format in custom dataset
|
il y a 1 an |
 Hamid Shojanazeri
|
df03fd4b12
Recipe to add a new language to Llama2 (#429)
|
il y a 1 an |
 Rahul A R
|
cc1029bcf1
update wordlist.txt
|
il y a 1 an |
|
Rahul A R
|
664429b726
Merge branch 'main' of github.com:rahul-sarvam/llama-recipes
|
il y a 1 an |
|
Rahul A R
|
e98f6de80d
typo
|
il y a 1 an |
 Kai Wu
|
03f1ca7817
fixed some typo to pass spellcheck
|
il y a 1 an |
|
Matthias Reso
|
83fae41195
Add test for chat completion formatting
|
il y a 1 an |
|
Kai Wu
|
d0b7a20c89
finetuning readme updated
|
il y a 1 an |
|
Hamid Shojanazeri
|
e3d750f49f
Update wordlist.txt
|
il y a 1 an |
|
Rahul A R
|
09028bf893
addressing Hamid's comments
|
il y a 1 an |
|
Matthias Reso
|
6d9d48d619
Use apply_chat_template instead of custom functions
|
il y a 1 an |
|
Matthias Reso
|
5efea160a2
Adapt test_finetuning to new model
|
il y a 1 an |
 rahul-sarvam
|
eb7ef4225f
Update recipes/multilingual/README.md
|
il y a 1 an |
|
rahul-sarvam
|
f1f335a591
Update recipes/multilingual/README.md
|
il y a 1 an |
|
rahul-sarvam
|
47556ce0a6
Update recipes/multilingual/README.md
|
il y a 1 an |
|
Matthias Reso
|
739483f262
Adjust test_grammar_datasets to stable sort
|
il y a 1 an |
|
Matthias Reso
|
b96e435cda
Adjust test_samsum_dataset to second model
|
il y a 1 an |
|
Matthias Reso
|
fac41298b0
Adapt test_custom_dataset to new model
|
il y a 1 an |
|
Matthias Reso
|
960014a3bb
Fix test_custom_dataset by introducing a stable sort algorithm
|
il y a 1 an |
|
Matthias Reso
|
b5583b31d5
Adapt test_grammar_dataset to new model
|
il y a 1 an |
|
Matthias Reso
|
17a6d16289
Test batching for both llama versions
|
il y a 1 an |
|
Kai Wu
|
7b1a9413d2
fixed a typo
|
il y a 1 an |
|
Kai Wu
|
41434dc825
formatted and removed duplicated or unused function get_total_flops() and byte2mb()
|
il y a 1 an |
|
Kai Wu
|
f2e80bae22
created a FlopMeasure class on top of FlopCounterMode instead of keep of copy of our own tflop_counter.py
|
il y a 1 an |
|
Matthias Reso
|
a414ca6a57
Update chat format for llama3
|
il y a 1 an |
|
Kai Wu
|
69e46887b4
handling incorrect profiling early stop caused by max_train_steps and add profiler.step() for each train step
|
il y a 1 an |
|
Matthias Reso
|
113ea18bf1
Replace LlamaTokenizer with AutoTokenizer
|
il y a 1 an |
 Beto
|
5979dbe996
Merging local with remote
|
il y a 1 an |
|
Kai Wu
|
34e0bf4c6e
second draft of this feature, seems to be working now
|
il y a 1 an |
|
Beto
|
d4cbfa1cc1
Merging upstream llama-recipes to current repo
|
il y a 1 an |
