better formating

recipes/fromSimPO/Qwen3-4B-fourchan.yaml

@@ -0,0 +1,50 @@
+# Model arguments
+model_name_or_path: Qwen/Qwen3-4B-Base
+model_revision: main
+torch_dtype: bfloat16
+attn_implementation: flash_attention_2
+
+# Data training arguments
+tokenizer_name_or_path: Qwen/Qwen3-4B # Custom tokenizer with <|im_start|> and <|im_end|> tokens
+# chat_template: "{% for message in messages %}\n{% if message['role'] == 'user' %}\n{{ '<|user|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'system' %}\n{{ '<|system|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'assistant' %}\n{{ '<|assistant|>\n'  + message['content'] + eos_token }}\n{% endif %}\n{% if loop.last and add_generation_prompt %}\n{{ '<|assistant|>' }}\n{% endif %}\n{% endfor %}"
+dataset_mixer:
+  wassname/v2ray_4chan_formatted: 0.8
+  wassname/ultrachat_200k_filtered: 0.2
+dataset_splits:
+- train_sft
+- test_sft
+preprocessing_num_workers: 12
+
+# SFT trainer config
+bf16: true
+do_eval: true
+eval_strategy: steps
+eval_steps: 200
+gradient_accumulation_steps: 32
+gradient_checkpointing: true
+gradient_checkpointing_kwargs:
+  use_reentrant: False
+hub_model_id: Qwen3-4B-4chan
+hub_strategy: every_save
+learning_rate: 2.0e-04
+log_level: info
+logging_steps: 5  
+logging_strategy: steps
+lr_scheduler_type: cosine
+max_seq_length: 2048
+max_steps: -1
+num_train_epochs: 3
+output_dir: /workspace/checkpoints_new/Qwen3-4B-4chan
+run_name: Qwen3-4B-4chan
+overwrite_output_dir: true
+per_device_eval_batch_size: 3
+per_device_train_batch_size: 2
+push_to_hub: false
+remove_unused_columns: true
+report_to:
+- wandb
+save_strategy: "steps"
+save_steps: 1000000
+save_total_limit: 1
+seed: 42
+warmup_ratio: 0.1

recipes/fromSimPO/SmolLM2-1.7B-sft.yaml

@@ -0,0 +1,49 @@
+# Model arguments
+model_name_or_path: HuggingFaceTB/SmolLM2-1.7B
+model_revision: main
+torch_dtype: bfloat16
+attn_implementation: flash_attention_2
+
+# Data training arguments
+tokenizer_name_or_path: HuggingFaceTB/SmolLM2-1.7B-Instruct # Custom tokenizer with <|im_start|> and <|im_end|> tokens
+# chat_template: "{% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{% if add_generation_prompt %}{{ '<|start_header_id|>assistant<|end_header_id|>\n\n' }}{% endif %}"
+dataset_mixer:
+  wassname/ultrachat_200k_filtered: 1.0
+dataset_splits:
+- train_sft
+- test_sft
+preprocessing_num_workers: 12
+
+# SFT trainer config
+bf16: true
+do_eval: true
+eval_strategy: steps
+eval_steps: 200
+gradient_accumulation_steps: 32
+gradient_checkpointing: true
+gradient_checkpointing_kwargs:
+  use_reentrant: False
+hub_model_id: SmolLM2-1.7B-sft
+hub_strategy: every_save
+learning_rate: 2.0e-04
+log_level: info
+logging_steps: 5  
+logging_strategy: steps
+lr_scheduler_type: cosine
+max_seq_length: 2048
+max_steps: -1
+num_train_epochs: 3
+output_dir: /workspace/checkpoints_new/SmolLM2-1.7B-sft
+run_name: SmolLM2-1.7B-sft
+overwrite_output_dir: true
+per_device_eval_batch_size: 4
+per_device_train_batch_size: 4
+push_to_hub: false
+remove_unused_columns: true
+report_to:
+- wandb
+save_strategy: "steps"
+save_steps: 1000000
+save_total_limit: 1
+seed: 42
+warmup_ratio: 0.1