Update cfg.yaml

cfg.yaml

@@ -28,7 +28,7 @@ dataset:
     text_answer_separator: <|answer|>
     text_prompt_start: <|prompt|>
     train_dataframe: data/user/oasst/train_full_allrank.pq
-    validation_dataframe: data/user/oasst/gpt4_val_v0.csv
+    validation_dataframe: data/user/oasst/val.csv
     validation_size: 0.01
     validation_strategy: custom
 environment:
@@ -39,7 +39,7 @@ environment:
     - '1'
     - '2'
     - '3'
-    huggingface_branch: refs/pr/8
+    huggingface_branch: main
     mixed_precision: true
     number_of_workers: 8
     seed: -1
@@ -47,10 +47,6 @@ environment:
     use_fsdp: false
 experiment_name: h2ogpt-gm-oasst1-en-xgen-7b-8k
 llm_backbone: Salesforce/xgen-7b-8k-base
-logging:
-    logger: Neptune
-    neptune_project: Zoo/h2o-llm
-    number_of_texts: 10
 output_directory: output/user/h2ogpt-gm-oasst1-en-xgen-7b-8k/
 prediction:
     batch_size_inference: 0
@@ -105,7 +101,6 @@ training:
     ppo_clip_value: 0.2
     ppo_epochs: 4
     ppo_generate_temperature: 1.0
-    reward_model: OpenAssistant/reward-model-deberta-v3-large-v2
     save_best_checkpoint: false
     scaling_factor_value_loss: 0.1
     schedule: Cosine