Skip to content

Fix SFT notebook parameter names for current API (RHOAIENG-37015) #739

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
openshift-merge-bot merged 4 commits into from
Apr 10, 2026
Merged

Fix SFT notebook parameter names for current API (RHOAIENG-37015) #739

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
b1d3a0d
fix: update SFT notebook to use new API parameter names (RHOAIENG-37015)
Apr 10, 2026
26f6293
Add newline at end of file
hrathina Apr 10, 2026
8b01a3f
Merge branch 'main' into fix/RHOAIENG-37015-tensorboard-pytorch28
hrathina Apr 10, 2026
989d849
Merge branch 'main' into fix/RHOAIENG-37015-tensorboard-pytorch28
hrathina Apr 10, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 4 additions & 5 deletions examples/kfto-sft-llm/sft.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -42,9 +42,9 @@
"# Model\n",
"model_name_or_path: Meta-Llama/Meta-Llama-3.1-8B-Instruct\n",
"model_revision: main\n",
"torch_dtype: bfloat16\n",
"dtype: bfloat16\n",
"attn_implementation: flash_attention_2 # one of eager (default), sdpa or flash_attention_2\n",
"use_liger: false # use Liger kernels\n",
"use_liger_kernel: false # use Liger kernels\n",
"\n",
"# PEFT / LoRA\n",
"use_peft: true\n",
Expand All @@ -69,8 +69,7 @@
" append_concat_token: false # add additional separator token\n",
"\n",
"# SFT\n",
"max_seq_length: 1024 # max sequence length for model and packing of the dataset\n",
"dataset_batch_size: 1000 # samples to tokenize per batch\n",
"max_length: 1024 # max sequence length for model and packing of the dataset\n",
"packing: false\n",
"\n",
"# Training\n",
Expand Down Expand Up @@ -175,7 +174,7 @@
" revision=model_args.model_revision,\n",
" trust_remote_code=model_args.trust_remote_code,\n",
" attn_implementation=model_args.attn_implementation,\n",
" torch_dtype=model_args.torch_dtype,\n",
" dtype=model_args.dtype,\n",
" use_cache=False if training_args.gradient_checkpointing or\n",
" training_args.fsdp_config.get(\"activation_checkpointing\",\n",
" False) else True,\n",
Expand Down