Fix total_mem → total_memory in V4 notebook (PyTorch API)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

notebooks/v4_instruct_grpo.ipynb

@@ -27,7 +27,7 @@
    "execution_count": null,
    "metadata": {},
    "outputs": [],
-   "source": "import torch\n\nprint(f\"CUDA available: {torch.cuda.is_available()}\")\nprint(f\"GPU: {torch.cuda.get_device_name(0)}\")\nprint(f\"VRAM: {torch.cuda.get_device_properties(0).total_mem / 1e9:.1f} GB\")\nprint(f\"bf16 support: {torch.cuda.is_bf16_supported()}\")\n\nfrom unsloth import FastLanguageModel\nprint(f\"\\n✓ Unsloth loaded\")\n\nimport trl\nassert trl.__version__ == \"0.24.0\", f\"Expected TRL 0.24.0, got {trl.__version__}\"\nprint(f\"✓ TRL {trl.__version__}\")\n\nimport transformers\nprint(f\"✓ Transformers {transformers.__version__}\")"
+   "source": "import torch\n\nprint(f\"CUDA available: {torch.cuda.is_available()}\")\nprint(f\"GPU: {torch.cuda.get_device_name(0)}\")\nprint(f\"VRAM: {torch.cuda.get_device_properties(0).total_memory / 1e9:.1f} GB\")\nprint(f\"bf16 support: {torch.cuda.is_bf16_supported()}\")\n\nfrom unsloth import FastLanguageModel\nprint(f\"\\n✓ Unsloth loaded\")\n\nimport trl\nassert trl.__version__ == \"0.24.0\", f\"Expected TRL 0.24.0, got {trl.__version__}\"\nprint(f\"✓ TRL {trl.__version__}\")\n\nimport transformers\nprint(f\"✓ Transformers {transformers.__version__}\")"
   },
   {
    "cell_type": "markdown",
@@ -135,7 +135,7 @@
    "execution_count": null,
    "metadata": {},
    "outputs": [],
-   "source": "from trl import GRPOConfig, GRPOTrainer\n\nFastLanguageModel.for_training(model)\n\nsmoke_config = GRPOConfig(\n    output_dir=str(CHECKPOINT_DIR / \"smoke\"),\n    num_generations=NUM_GENERATIONS,\n    scale_rewards=SCALE_REWARDS,\n    max_completion_length=MAX_COMPLETION_LENGTH,\n    max_steps=1,\n    temperature=TEMPERATURE,\n    beta=BETA,\n    per_device_train_batch_size=BATCH_SIZE,\n    gradient_accumulation_steps=1,\n    learning_rate=LEARNING_RATE,\n    fp16=False,\n    bf16=True,\n    logging_steps=1,\n    save_steps=999,\n    report_to=\"none\",\n    max_prompt_length=MAX_SEQ_LENGTH // 2,\n    seed=42,\n    remove_unused_columns=False,\n)\n\n\nclass UnslothGRPOTrainer(GRPOTrainer):\n    \"\"\"Wraps generation with Unsloth for_inference()/for_training().\"\"\"\n    def _generate(self, prompts, images):\n        FastLanguageModel.for_inference(self.model)\n        try:\n            result = super()._generate(prompts, images)\n        finally:\n            FastLanguageModel.for_training(self.model)\n        return result\n\n\nsmoke_trainer = UnslothGRPOTrainer(\n    model=model,\n    reward_funcs=commerce_reward_fn,\n    args=smoke_config,\n    train_dataset=train_dataset,\n    processing_class=tokenizer,\n)\n\nt0 = time.time()\nsmoke_trainer.train()\nstep_time = time.time() - t0\n\npeak_vram = torch.cuda.max_memory_allocated() / 1e9\nprint(f\"\\n✓ Smoke test passed!\")\nprint(f\"  Step time: {step_time:.0f}s\")\nprint(f\"  Peak VRAM: {peak_vram:.1f}GB / {torch.cuda.get_device_properties(0).total_mem / 1e9:.1f}GB\")\nprint(f\"  Estimated full run ({MAX_STEPS} steps): {step_time * MAX_STEPS / 3600:.1f}h\")\n\ndel smoke_trainer\ngc.collect(); torch.cuda.empty_cache()"
+   "source": "from trl import GRPOConfig, GRPOTrainer\n\nFastLanguageModel.for_training(model)\n\nsmoke_config = GRPOConfig(\n    output_dir=str(CHECKPOINT_DIR / \"smoke\"),\n    num_generations=NUM_GENERATIONS,\n    scale_rewards=SCALE_REWARDS,\n    max_completion_length=MAX_COMPLETION_LENGTH,\n    max_steps=1,\n    temperature=TEMPERATURE,\n    beta=BETA,\n    per_device_train_batch_size=BATCH_SIZE,\n    gradient_accumulation_steps=1,\n    learning_rate=LEARNING_RATE,\n    fp16=False,\n    bf16=True,\n    logging_steps=1,\n    save_steps=999,\n    report_to=\"none\",\n    max_prompt_length=MAX_SEQ_LENGTH // 2,\n    seed=42,\n    remove_unused_columns=False,\n)\n\n\nclass UnslothGRPOTrainer(GRPOTrainer):\n    \"\"\"Wraps generation with Unsloth for_inference()/for_training().\"\"\"\n    def _generate(self, prompts, images):\n        FastLanguageModel.for_inference(self.model)\n        try:\n            result = super()._generate(prompts, images)\n        finally:\n            FastLanguageModel.for_training(self.model)\n        return result\n\n\nsmoke_trainer = UnslothGRPOTrainer(\n    model=model,\n    reward_funcs=commerce_reward_fn,\n    args=smoke_config,\n    train_dataset=train_dataset,\n    processing_class=tokenizer,\n)\n\nt0 = time.time()\nsmoke_trainer.train()\nstep_time = time.time() - t0\n\npeak_vram = torch.cuda.max_memory_allocated() / 1e9\nprint(f\"\\n✓ Smoke test passed!\")\nprint(f\"  Step time: {step_time:.0f}s\")\nprint(f\"  Peak VRAM: {peak_vram:.1f}GB / {torch.cuda.get_device_properties(0).total_memory / 1e9:.1f}GB\")\nprint(f\"  Estimated full run ({MAX_STEPS} steps): {step_time * MAX_STEPS / 3600:.1f}h\")\n\ndel smoke_trainer\ngc.collect(); torch.cuda.empty_cache()"
   },
   {
    "cell_type": "markdown",