Remove tokenizer parameter from SFTTrainer instantiation

Author: qgallouedec

notebooks/1/4.ipynb

@@ -1944,7 +1944,7 @@
     },
     {
       "cell_type": "code",
-      "execution_count": 36,
+      "execution_count": null,
       "metadata": {
         "colab": {
           "base_uri": "https://localhost:8080/",
@@ -2384,7 +2384,6 @@
         "    model=model,\n",
         "    args=sft_config,\n",
         "    train_dataset=train_dataset,\n",
-        "    tokenizer=tokenizer,\n",
         "    peft_config=peft_config,\n",
         "    packing=True,\n",
         "    dataset_kwargs={\n",

units/en/unit1/3.md

@@ -289,7 +289,6 @@ from trl import SFTTrainer
 
 trainer = SFTTrainer(
     model=model,
-    tokenizer=tokenizer,
     train_dataset=dataset["train"],
     args=config,
 )
@@ -504,7 +503,6 @@ config = SFTConfig(
 # Train!
 trainer = SFTTrainer(
     model=model,
-    tokenizer=tokenizer,
     train_dataset=dataset["train"],
     args=config,
 )

units/en/unit1/3a.md

@@ -142,7 +142,6 @@ trainer = SFTTrainer(
     model=model,
     args=SFTConfig(output_dir="lora-adapter", num_train_epochs=1, per_device_train_batch_size=2),
     train_dataset=dataset["train"],
-    tokenizer=tokenizer,
     peft_config=peft_config,
     packing=True,
 )

units/en/unit1/4.md

@@ -1186,7 +1186,6 @@ from trl import SFTTrainer
 
 lora_trainer = SFTTrainer(
     model=model,
-    tokenizer=tokenizer,
     train_dataset=formatted_dataset,  # dataset with a "text" field or messages + dataset_text_field in config
     args=training_config,
     peft_config=peft_config,  # << enable LoRA
@@ -1204,7 +1203,6 @@ We instantiate the trainer, capture a pre-training baseline generation, launch `
 
 trainer = SFTTrainer(
     model=model,
-    tokenizer=tokenizer,
     train_dataset=dataset["train"],
     args=config,
 )

units/en/unit1/5.md

@@ -64,7 +64,6 @@ config = SFTConfig(
 # Train
 trainer = SFTTrainer(
     model=model,
-    tokenizer=tokenizer,
     train_dataset=dataset["train"],
     args=config,
 )

v1/3_parameter_efficient_finetuning/notebooks/finetune_sft_peft.ipynb

@@ -279,7 +279,6 @@
     "    train_dataset=dataset[\"train\"],\n",
     "    peft_config=peft_config,  # LoRA configuration\n",
     "    max_seq_length=max_seq_length,  # Maximum sequence length\n",
-    "    tokenizer=tokenizer,\n",
     "    packing=True,  # Enable input packing for efficiency\n",
     "    dataset_kwargs={\n",
     "        \"add_special_tokens\": False,  # Special tokens handled by template\n",

v1/5_vision_language_models/notebooks/vlm_sft_sample.ipynb

@@ -369,7 +369,6 @@
     "    eval_dataset=ds[\"test\"],\n",
     "    data_collator=collate_fn,\n",
     "    peft_config=peft_config,\n",
-    "    tokenizer=processor.tokenizer,\n",
     ")\n",
     "\n",
     "# TODO: 🦁 🐕 align the SFTTrainer params with your chosen dataset. For example, if you are using the `bigcode/the-stack-smol` dataset, you will need to choose the `content` column`"

v1/es/1_instruction_tuning/notebooks/sft_finetuning_example.ipynb

@@ -168,7 +168,6 @@
     "    model=model,\n",
     "    args=sft_config,\n",
     "    train_dataset=ds[\"train\"],\n",
-    "    tokenizer=tokenizer,\n",
     "    eval_dataset=ds[\"test\"],\n",
     ")\n",
     "\n",

v1/es/3_parameter_efficient_finetuning/notebooks/finetune_sft_peft.ipynb

@@ -279,7 +279,6 @@
     "    train_dataset=dataset[\"train\"],\n",
     "    peft_config=peft_config,  # Configuración LoRA\n",
     "    max_seq_length=max_seq_length,  # Longitud máxima de la secuencia\n",
-    "    tokenizer=tokenizer,\n",
     "    packing=True,  # Habilita el empaquetado de entrada para mayor eficiencia\n",
     "    dataset_kwargs={\n",
     "        \"add_special_tokens\": False,  # Los tokens especiales son manejados por la plantilla\n",

v1/ja/1_instruction_tuning/notebooks/chat_templates_example.ipynb

@@ -168,7 +168,6 @@
     "    model=model,\n",
     "    args=sft_config,\n",
     "    train_dataset=ds[\"train\"],\n",
-    "    tokenizer=tokenizer,\n",
     "    eval_dataset=ds[\"test\"],\n",
     ")\n",
     "\n",