salman 54dd7abfc1 Process reward models (#2241) ...

* adding model_cfg to set num_labels

* using a num_labels field instead

* linting

* WIP stepwise prompt tokenizer

* this should work?

* trainer working?

* pushing to runpod

* fixing saving

* updating conf

* updating config, adding docs

* adding stepwise supervision docpage

* updating tests

* adding test for dataset

* fixing tests

* linting

* addressing some comments

* adding additional cfg fields support

* updating tests, fixing cfg

* fixing tests

* updating loss

* Update test_process_reward_model_smollm2.py

* updating loss values and seed

* dumb pre-commit

2025-01-29 00:08:33 -05:00

..

messages

wip add new proposed message structure (#1904)

2024-10-13 12:15:18 -04:00

__init__.py

move shared pytest conftest to top level tests (#2099) [skip ci]

2024-11-22 15:05:42 -05:00

conftest.py

fix: use apply_chat_template to find turn boundaries and allow tool_calling field (#2179) [skip ci]

2024-12-17 16:42:21 -05:00

test_alpaca.py

fix broken linting (#1541)

2024-04-19 01:03:04 -04:00

test_chat_template_utils.py

Feat: Add support for tokenizer’s or custom jinja chat_template (#1970)

2024-10-29 10:14:51 +07:00

test_chat_templates_advanced.py

fix: use apply_chat_template to find turn boundaries and allow tool_calling field (#2179) [skip ci]

2024-12-17 16:42:21 -05:00

test_chat_templates.py

fix: use apply_chat_template to find turn boundaries and allow tool_calling field (#2179) [skip ci]

2024-12-17 16:42:21 -05:00

test_dpo_chat_templates.py

Feat: Add support for tokenizer’s or custom jinja chat_template (#1970)

2024-10-29 10:14:51 +07:00

test_raw_io.py

fix broken linting (#1541)

2024-04-19 01:03:04 -04:00

test_stepwise.py

Process reward models (#2241)

2025-01-29 00:08:33 -05:00