Update README.md

README.md

@@ -216,13 +216,19 @@ a:hover {text-decoration: underline;}
       <p>The model then went through DPO training using approx 1100 chosen examples from the SFT dataset that were of exceptional quality or showed verifiable instruction following. Rejected samples were generated using another Llama 3.3 finetune that is known for poor instruction following.</p>
       <h3 class="subheading">SFT 1*H200</h3>
       <div class="data-box">
-        <pre style="overflow-x: auto; color: #e1e9f0; margin: 0;">base_model: zerofata/L3.3-GeneticLemonade-Unleashed-70B
+        <pre style="overflow-x: auto; color: #e1e9f0; margin: 0;"># ====================
+# MODEL CONFIGURATION
+# ====================
+base_model: zerofata/L3.3-GeneticLemonade-Unleashed-70B
 model_type: AutoModelForCausalLM
 tokenizer_type: AutoTokenizer
 special_tokens:
   pad_token: "<|finetune_right_pad_id|>"
 chat_template: llama3
 
+# ====================
+# DATASET CONFIGURATION
+# ====================
 datasets:
   - path: ./dataset.jsonl
     type: chat_template
@@ -237,14 +243,37 @@ datasets:
       assistant: ["assistant"]
       system: ["system"]
 
-train_on_inputs: false
+test_datasets:
+  - path: ./validate_dataset.jsonl
+    type: chat_template
+    split: train
+    chat_template_strategy: tokenizer
+    field_messages: messages
+    message_property_mappings:
+      role: role
+      content: content
+    roles:
+      user: ["user"]
+      assistant: ["assistant"]
+      system: ["system"]
+
+dataset_prepared_path:
+train_on_inputs: false  # Only train on assistant responses
+
+# ====================
+# QLORA CONFIGURATION
+# ====================
 adapter: qlora
 load_in_4bit: true
 lora_r: 64
 lora_alpha: 128
 lora_dropout: 0.1
 lora_target_linear: true
+# lora_modules_to_save:  # Uncomment only if you added NEW tokens
 
+# ====================
+# TRAINING PARAMETERS
+# ====================
 num_epochs: 2
 micro_batch_size: 4
 gradient_accumulation_steps: 2
@@ -255,25 +284,68 @@ warmup_ratio: 0.05
 weight_decay: 0.01
 max_grad_norm: 1.0
 
+# ====================
+# SEQUENCE & PACKING
+# ====================
 sequence_len: 8192
 sample_packing: true
+eval_sample_packing: false
 pad_to_sequence_len: true
 
+# ====================
+# HARDWARE OPTIMIZATIONS
+# ====================
 bf16: auto
 flash_attention: true
-gradient_checkpointing: true</pre>
+gradient_checkpointing: true
+
+# ====================
+# EVALUATION & CHECKPOINTING
+# ====================
+evaluation_strategy: steps
+eval_steps: 5
+save_strategy: steps
+save_steps: 5
+save_total_limit: 5  # Keep best + last few checkpoints
+load_best_model_at_end: true
+metric_for_best_model: eval_loss
+greater_is_better: false
+early_stopping_patience: 5
+
+# ====================
+# LOGGING & OUTPUT
+# ====================
+output_dir: ./output_model
+logging_steps: 2
+save_safetensors: true
+
+# ====================
+# WANDB TRACKING
+# ====================
+wandb_project: project_name
+# wandb_entity: your_entity  # Uncomment and set if needed
+# wandb_name: your_run_name  # Uncomment and set if needed</pre>
       </div>
       <h3 class="subheading">DPO 2*H200</h3>
       <div class="data-box">
-        <pre style="overflow-x: auto; color: #e1e9f0; margin: 0;">base_model: ApocalypseParty/unleashed-fulldata30
+        <pre style="overflow-x: auto; color: #e1e9f0; margin: 0;"># ====================
+# MODEL CONFIGURATION
+# ====================
+base_model: ApocalypseParty/unleashed-fulldata30
 model_type: AutoModelForCausalLM
 tokenizer_type: AutoTokenizer
 special_tokens: {}
 chat_template: tokenizer_default
 
+# ====================
+# RL/DPO CONFIGURATION
+# ====================
 rl: dpo
 rl_beta: 0.07
 
+# ====================
+# DATASET CONFIGURATION
+# ====================
 datasets:
   - path: ./dpo_cleaned-v3_deduplicated.jsonl
     type: chat_template.default
@@ -287,15 +359,23 @@ datasets:
       system: ["system"]
       user: ["user"]
       assistant: ["assistant"]
+dataset_prepared_path:
+train_on_inputs: false  # Only train on assistant responses
 
-train_on_inputs: false
+# ====================
+# QLORA CONFIGURATION
+# ====================
 adapter: qlora
 load_in_4bit: true
 lora_r: 32
 lora_alpha: 64
 lora_dropout: 0.05
 lora_target_linear: true
+# lora_modules_to_save:  # Uncomment only if you added NEW tokens
 
+# ====================
+# TRAINING PARAMETERS
+# ====================
 num_epochs: 1
 micro_batch_size: 4
 gradient_accumulation_steps: 2
@@ -306,17 +386,46 @@ warmup_steps: 5
 weight_decay: 0.01
 max_grad_norm: 1.0
 
+# ====================
+# SEQUENCE CONFIGURATION
+# ====================
 sequence_len: 4096
 pad_to_sequence_len: true
 
+# ====================
+# HARDWARE OPTIMIZATIONS
+# ====================
 bf16: auto
 tf32: false
 flash_attention: true
 gradient_checkpointing: offload
-deepspeed: deepspeed_configs/zero1.json</pre>
+deepspeed: deepspeed_configs/zero1.json
+
+# ====================
+# CHECKPOINTING
+# ====================
+save_steps: 10
+save_total_limit: 10
+load_best_model_at_end: true
+metric_for_best_model: eval_loss
+greater_is_better: false
+
+# ====================
+# LOGGING & OUTPUT
+# ====================
+output_dir: ./dpo_model
+logging_steps: 2
+save_safetensors: true
+
+# ====================
+# WANDB TRACKING
+# ====================
+wandb_project: project_name
+# wandb_entity: your_entity  # Uncomment and set if needed
+# wandb_name: your_run_name  # Uncomment and set if needed</pre>
       </div>
     </div>
   </div>
 </div>
 </body>
-</html>
+</html>