set torchao quant config on config.json of saved model

* add validation around cce + chunked_ce

* return on end of validation method

.coderabbit.yaml

@@ -0,0 +1,16 @@
+# yaml-language-server: $schema=https://coderabbit.ai/integrations/schema.v2.json
+language: "en-US"
+early_access: false
+reviews:
+  profile: "chill"
+  request_changes_workflow: false
+  high_level_summary: true
+  review_status: true
+  collapse_walkthrough: true
+  poem: false
+  sequence_diagrams: false
+  auto_review:
+    enabled: true
+    drafts: false
+chat:
+  auto_reply: true

.github/workflows/main.yml

@@ -87,7 +87,6 @@ jobs:
             python_version: "3.11"
             pytorch: 2.6.0
             axolotl_extras:
-            is_latest: true
           - cuda: 126
             cuda_version: 12.6.3
             python_version: "3.11"
@@ -98,6 +97,7 @@ jobs:
             python_version: "3.11"
             pytorch: 2.7.1
             axolotl_extras:
+            is_latest: true
           - cuda: 128
             cuda_version: 12.8.1
             python_version: "3.11"

.github/workflows/tests-nightly.yml

@@ -106,6 +106,13 @@ jobs:
             num_gpus: 1
             axolotl_extras:
             nightly_build: "true"
+          - cuda: 126
+            cuda_version: 12.6.3
+            python_version: "3.11"
+            pytorch: 2.7.1
+            num_gpus: 1
+            axolotl_extras:
+            nightly_build: "true"
     steps:
       - name: Checkout
         uses: actions/checkout@v4
@@ -130,3 +137,45 @@ jobs:
       - name: Run tests job on Modal
         run: |
           modal run cicd.e2e_tests
+  docker-e2e-multigpu-tests:
+    if: github.repository_owner == 'axolotl-ai-cloud'
+    # this job needs to be run on self-hosted GPU runners...
+    runs-on: [self-hosted, modal]
+    timeout-minutes: 120
+    needs: [pre-commit, pytest, docker-e2e-tests]
+
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          - cuda: 126
+            cuda_version: 12.6.3
+            python_version: "3.11"
+            pytorch: 2.7.1
+            num_gpus: 2
+            axolotl_extras:
+            nightly_build: "true"
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+      - name: Install Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.11"
+      - name: Install Modal
+        run: |
+          python -m pip install --upgrade pip
+          pip install modal==1.0.2 jinja2
+      - name: Update env vars
+        run: |
+          echo "BASE_TAG=main-base-py${{ matrix.python_version }}-cu${{ matrix.cuda }}-${{ matrix.pytorch }}" >> $GITHUB_ENV
+          echo "PYTORCH_VERSION=${{ matrix.pytorch}}" >> $GITHUB_ENV
+          echo "AXOLOTL_ARGS=${{ matrix.axolotl_args}}" >> $GITHUB_ENV
+          echo "AXOLOTL_EXTRAS=${{ matrix.axolotl_extras}}" >> $GITHUB_ENV
+          echo "CUDA=${{ matrix.cuda }}" >> $GITHUB_ENV
+          echo "N_GPUS=${{ matrix.num_gpus }}" >> $GITHUB_ENV
+          echo "NIGHTLY_BUILD=${{ matrix.nightly_build }}" >> $GITHUB_ENV
+          echo "CODECOV_TOKEN=${{ secrets.CODECOV_TOKEN }}" >> $GITHUB_ENV
+      - name: Run tests job on Modal
+        run: |
+          modal run cicd.multigpu

codecov.yml

@@ -22,6 +22,7 @@ coverage:
         only_pulls: true
         flags: null
         paths: null
+        informational: true
     patch:
       default:
         # basic

requirements.txt

@@ -15,7 +15,7 @@ huggingface_hub>=0.33.0
 peft==0.16.0
 transformers==4.53.2
 tokenizers>=0.21.1
-accelerate==1.8.1
+accelerate==1.9.0
 datasets==4.0.0
 deepspeed>=0.17.0
 trl==0.19.1

src/axolotl/cli/quantize.py

@@ -43,7 +43,7 @@ def do_quantize(
             "No quantization configuration found. Please specify either qat or quantization in your config file."
         )
 
-    model_path = cli_args.get("model_path") or cfg.output_dir
+    model_path = cli_args.get("base_model") or cfg.output_dir
     if weight_dtype := cli_args.get("weight_dtype"):
         weight_dtype = TorchIntDType[weight_dtype]
     else:

src/axolotl/core/datasets/chat.py

@@ -2,7 +2,6 @@
 chat dataset module
 """
 
-import os
 from typing import Callable, Optional, Union
 
 from datasets import Dataset
@@ -41,14 +40,10 @@ class TokenizedChatDataset(Dataset):
                 )
             return ex.tokenized(model_transform)
 
-        process_or_cpu_count: int = (
-            process_count or os.cpu_count()  # type: ignore[assignment]
-        )
-        num_proc = min(32, process_or_cpu_count)
         features = data.features.keys()
         tokenized_data = data.map(
             map_fn,
-            num_proc=num_proc,
+            num_proc=process_count,
             keep_in_memory=keep_in_memory,
             remove_columns=features,
             desc="Tokenizing Chats",

src/axolotl/core/trainers/grpo/__init__.py

@@ -148,7 +148,7 @@ class GRPOStrategy:
 
     @classmethod
     def get_blocklist_args_kwargs(cls) -> list[str]:
-        return ["dataset_num_proc", "max_length"]
+        return ["dataset_num_proc", "max_length", "include_tokens_per_second"]
 
     @classmethod
     def get_reward_func(cls, reward_func_fqn: str) -> RewardFunc:

src/axolotl/datasets.py

@@ -1,7 +1,5 @@
 """Module containing Dataset functionality"""
 
-import os
-
 import torch
 from datasets import Dataset, IterableDataset
 
@@ -46,7 +44,6 @@ class TokenizedPromptDataset(Dataset):
 
     def process(self, dataset):
         features = dataset.features.keys()
-        num_proc = min(64, self.process_count if self.process_count else os.cpu_count())
 
         map_kwargs = {}
         if self.prompt_tokenizer.supports_batched:
@@ -59,13 +56,13 @@ class TokenizedPromptDataset(Dataset):
         ):
             dataset = dataset.filter(
                 self.prompt_tokenizer.filter_rows,
-                num_proc=num_proc,
+                num_proc=self.process_count,
                 desc="Strategy Filtering Rows",
             )
 
         return dataset.map(
             self.prompt_tokenizer.tokenize_prompt,
-            num_proc=num_proc,
+            num_proc=self.process_count,
             remove_columns=features,
             keep_in_memory=self.keep_in_memory,
             desc="Tokenizing Prompts",

src/axolotl/integrations/cut_cross_entropy/args.py

@@ -41,3 +41,13 @@ class CutCrossEntropyArgs(BaseModel):
             )
 
         return data
+
+    @model_validator(mode="before")
+    @classmethod
+    def check_chunked_cross_entropy_not_set(cls, data):
+        if data.get("chunked_cross_entropy"):
+            raise ValueError(
+                "Cut Cross Entropy does not support chunked cross entropy. "
+                "Please set `chunked_cross_entropy` to `False` or disable Cut Cross Entropy."
+            )
+        return data

src/axolotl/loaders/model.py

@@ -163,15 +163,6 @@ class ModelLoader:
         # Build the model
         PLUGIN_MANAGER.pre_model_load(self.cfg)
         skip_move_to_device = self._build_model()
-        
-        # Check if the model is a GraniteConfig object
-        if hasattr(self, 'model') and self.model.__class__.__name__ == "GraniteConfig":
-            LOG.error("The model loaded is a GraniteConfig object, not a proper model.")
-            LOG.error("This is likely because the model type 'GraniteConfig' is not supported.")
-            LOG.error("Please use a different model type or ensure the model is properly configured.")
-            LOG.error("Setting trust_remote_code=True might help if the model requires custom code.")
-            raise ValueError("Model loaded is a GraniteConfig object, not a proper model. Use a supported model type or set trust_remote_code=True.")
-            
         PLUGIN_MANAGER.post_model_build(self.cfg, self.model)
 
         # Post-build model configuration
@@ -225,27 +216,15 @@ class ModelLoader:
 
     def _resize_token_embeddings(self):
         """Resize token embeddings if needed."""
-        # Skip if model doesn't have the necessary methods
-        if not hasattr(self.model, "get_input_embeddings"):
-            LOG.warning("Model does not have get_input_embeddings method, skipping token embedding resize")
-            return
-            
-        # Check if get_input_embeddings returns None
-        input_embeddings = self.model.get_input_embeddings()
-        if input_embeddings is None:
-            LOG.warning("Model's get_input_embeddings returned None, skipping token embedding resize")
-            return
-            
         embeddings_len = (
             math.ceil(len(self.tokenizer) / 32) * 32
             if self.cfg.resize_token_embeddings_to_32x
             else len(self.tokenizer)
         )
-        
+        if hasattr(self.model, "get_input_embeddings") and (
-        if hasattr(input_embeddings, "num_embeddings") and (
+            self.model.get_input_embeddings().num_embeddings < embeddings_len
-            input_embeddings.num_embeddings < embeddings_len
             or (
-                input_embeddings.num_embeddings > embeddings_len
+                self.model.get_input_embeddings().num_embeddings > embeddings_len
                 and self.cfg.shrink_embeddings
             )
         ):
@@ -254,24 +233,14 @@ class ModelLoader:
                 self.model_config.model_type != "llava"
             ):
                 resize_kwargs["mean_resizing"] = self.cfg.mean_resizing_embeddings
-                
-            if hasattr(self.model, "resize_token_embeddings"):
             self.model.resize_token_embeddings(embeddings_len, **resize_kwargs)
         else:
-                LOG.warning("Model does not have resize_token_embeddings method, skipping resize")
-        else:
-            if hasattr(self.model, "tie_weights"):
             self.model.tie_weights()
 
     def _adjust_model_config(self):
-        # Skip if model doesn't have config attribute
-        if not hasattr(self.model, "config"):
-            LOG.warning("Model does not have config attribute, skipping model config adjustments")
-            return
-            
-        # Handle max_position_embeddings
         if (
-            hasattr(self.model.config, "max_position_embeddings")
+            hasattr(self.model, "config")
+            and hasattr(self.model.config, "max_position_embeddings")
             and self.model.config.max_position_embeddings
             and self.cfg.sequence_len > self.model.config.max_position_embeddings
         ):
@@ -281,17 +250,17 @@ class ModelLoader:
             )
             self.model.config.max_position_embeddings = self.cfg.sequence_len
 
-        # Handle bos_token_id
         if (
-            hasattr(self.model.config, "bos_token_id")
+            hasattr(self.model, "config")
+            and hasattr(self.model.config, "bos_token_id")
             and self.model.config.bos_token_id
             and self.model.config.bos_token_id != self.tokenizer.bos_token_id
         ):
             self.model.config.bos_token_id = self.tokenizer.bos_token_id
 
-        # Handle eos_token_id
         if (
-            hasattr(self.model.config, "eos_token_id")
+            hasattr(self.model, "config")
+            and hasattr(self.model.config, "eos_token_id")
             and self.model.config.eos_token_id
             and self.model.config.eos_token_id != self.tokenizer.eos_token_id
         ):
@@ -323,12 +292,9 @@ class ModelLoader:
         if self.cfg.adapter in ["lora", "qlora"]:
             needs_fa2_dtype = True
             if self.cfg.gradient_checkpointing:
-                if hasattr(self.model, "gradient_checkpointing_enable"):
                 self.model.gradient_checkpointing_enable(
                     gradient_checkpointing_kwargs=self.cfg.gradient_checkpointing_kwargs
                 )
-                else:
-                    LOG.warning("Model does not have gradient_checkpointing_enable method, skipping gradient checkpointing")
 
         self._prepare_model_for_quantization()
 
@@ -405,14 +371,11 @@ class ModelLoader:
             self.model.is_parallelizable = True
             self.model.model_parallel = True
 
-        if hasattr(self.model, "named_parameters"):
         if not any(
             param.requires_grad
             for _, param in self.model.named_parameters(recurse=True)
         ):
             LOG.warning("There are no parameters that require gradient updates")
-        else:
-            LOG.warning("Model does not have named_parameters attribute, skipping gradient check")
 
         if self.cfg.flash_optimum:
             from optimum.bettertransformer import BetterTransformer
@@ -420,10 +383,7 @@ class ModelLoader:
             self.model = BetterTransformer.transform(self.model)
 
         if self.cfg.adapter is not None:
-            if hasattr(self.model, "device"):
             log_gpu_memory_usage(LOG, "after adapters", self.model.device)
-            else:
-                LOG.warning("Model does not have device attribute, skipping memory usage logging")
 
         for _ in range(3):
             gc.collect()
@@ -740,10 +700,6 @@ class ModelLoader:
             and self.model_type != "AutoModelForCausalLM"
             and not self.cfg.trust_remote_code
         ):
-            if self.model_type == "GraniteSpeechConfig" and not hasattr(self.model_config, 'vocab_size'):
-                # Set vocab_size from tokenizer or use a reasonable default
-                self.model_config.vocab_size = getattr(self.model_config, 'vocab_size', 50257)
-
             if self.cfg.gptq:
                 self.model = self.auto_model_loader.from_pretrained(
                     self.base_model,
@@ -751,21 +707,7 @@ class ModelLoader:
                     trust_remote_code=self.cfg.trust_remote_code or False,
                     **self.model_kwargs,
                 )
-
-            elif self.model_type == "GraniteSpeechConfig":
-                # Use the actual model class for Granite Speech
-                self.model = transformers.GraniteSpeechForCausalLM.from_pretrained(
-                    self.base_model,
-                    config=self.model_config,
-                    trust_remote_code=self.cfg.trust_remote_code or False,
-                    **self.model_kwargs,
-                )
-
             else:
-                if not hasattr(self.model_config, 'vocab_size'):
-                    LOG.warning("Model config does not have vocab_size attribute, setting to 50257")
-                    self.model_config.vocab_size = 50257
-
                 self.model = getattr(transformers, self.model_type).from_pretrained(
                     self.base_model,
                     config=self.model_config,
@@ -849,19 +791,13 @@ class ModelLoader:
         dest = {"dtype": dist_dtype}
         if self.cfg.lora_on_cpu:
             dest["device"] = "cpu"
-        
-        # Check if the model has named_modules attribute
-        if not hasattr(self.model, "named_modules"):
-            LOG.warning("Model does not have named_modules attribute, skipping embedding dtype conversion")
-            return
-            
         for name, module in self.model.named_modules():
             if "norm" in name:
                 module.to(dist_dtype)
             if before_kbit_train_or_finetune:
                 if name.endswith(".gate"):
                     module.to(dist_dtype)
-                if self.model_config.model_type == "btlm" and "lm_head" in name:
+                if self.model_config.model_type == "btlm":
                     # don't upcast lm_head for btlm
                     continue
             if any(m in name for m in embedding_modules) and hasattr(module, "weight"):

src/axolotl/loaders/tokenizer.py

@@ -188,7 +188,8 @@ def load_tokenizer(cfg: DictDefault) -> PreTrainedTokenizer:
         tokenizer.padding_side = "left"
 
     # Qwen base only has single token, so we need to set the special tokens
-    if cfg.is_qwen_derived_model:
+    # the following check is for Qwen1 base models
+    if cfg.is_qwen_derived_model and hasattr(tokenizer, "eod_id"):
         token_ids = ["bos_token_id", "eos_token_id", "pad_token_id", "unk_token_id"]
         for attr_name in token_ids:
             if getattr(tokenizer, attr_name) is None:

src/axolotl/logging_config.py

@@ -113,7 +113,7 @@ DEFAULT_LOGGING_CONFIG: Dict[str, Any] = {
     "loggers": {
         "axolotl": {
             "handlers": ["color_console"],
-            "level": os.getenv("AXOLOTL_LOG_LEVEL", DEFAULT_AXOLOTL_LOG_LEVEL),
+            "level": os.getenv("AXOLOTL_LOG_LEVEL", DEFAULT_AXOLOTL_LOG_LEVEL).upper(),
             "propagate": False,
         },
     },

src/axolotl/monkeypatch/lora_kernels.py

@@ -151,6 +151,11 @@ def get_attention_cls_from_config(cfg: DictDefault) -> Type[nn.Module]:
 
         return MllamaTextSelfAttention
 
+    if model_type == "llama4":
+        from transformers.models.llama4.modeling_llama4 import Llama4TextAttention
+
+        return Llama4TextAttention
+
     try:
         # Dynamically import the module and attention class
         module_path = f"transformers.models.{model_type}.modeling_{model_type}"

src/axolotl/train.py

@@ -80,15 +80,7 @@ def setup_model_and_tokenizer(
 
     model_loader = ModelLoader(cfg, tokenizer, processor=processor)
     model, peft_config = model_loader.load()
-    
+    if model.generation_config is not None:
-    # Check if model is actually a GraniteConfig object
-    if model.__class__.__name__ == "GraniteConfig":
-        LOG.error("The model loaded is a GraniteConfig object, not a proper model.")
-        LOG.error("This is likely because the model type 'GraniteConfig' is not supported.")
-        LOG.error("Please use a different model type or ensure the model is properly configured.")
-        raise ValueError("Model loaded is a GraniteConfig object, not a proper model. Use a supported model type.")
-        
-    if hasattr(model, "generation_config") and model.generation_config is not None:
         model.generation_config.do_sample = True
 
     # Apply freezing if specified
@@ -98,10 +90,7 @@ def setup_model_and_tokenizer(
             any(embed in param for embed in ["lm_head", "embed_tokens"])
             for param in cfg.unfrozen_parameters
         ):
-            if hasattr(model, "enable_input_require_grads"):
             model.enable_input_require_grads()
-            else:
-                LOG.warning("Model does not have enable_input_require_grads method, skipping")
 
     return model, tokenizer, peft_config, processor
 
@@ -257,12 +246,9 @@ def save_trained_model(
     LOG.info(f"Training completed! Saving trained model to {cfg.output_dir}.")
 
     # Post training module hooks
-    if hasattr(model, "named_modules"):
     for name, module in model.named_modules():
         if hasattr(module, "_post_training"):
             module._post_training(model, name)  # pylint: disable=protected-access
-    else:
-        LOG.warning("Model does not have named_modules attribute, skipping post training hooks")
 
     # handle QAT
     if cfg.qat:
@@ -322,17 +308,11 @@ def save_trained_model(
             model = BetterTransformer.reverse(model)
 
         if cfg.rl and cfg.adapter and not cfg.rl_adapter_ref_model:
-            if hasattr(trainer.model, "save_pretrained"):
             trainer.model.save_pretrained(
                 cfg.output_dir, safe_serialization=safe_serialization
             )
-            else:
-                LOG.warning("Trainer model does not have save_pretrained method, skipping save")
 
-        if hasattr(model, "save_pretrained"):
         model.save_pretrained(cfg.output_dir, safe_serialization=safe_serialization)
-        else:
-            LOG.warning("Model does not have save_pretrained method, skipping save")
 
     if hasattr(cfg, "llmcompressor") and cfg.llmcompressor:
         # TODO: add integration support so this can be implemented completely within the plugin
@@ -418,10 +398,7 @@ def save_initial_configs(
     tokenizer.save_pretrained(str(output_dir))
     if hasattr(model, "config"):
         LOG.info(f"Pre-saving model config to {cfg.output_dir}...")
-        if hasattr(model.config, "save_pretrained"):
         model.config.save_pretrained(str(output_dir))
-        else:
-            LOG.warning("Model config does not have save_pretrained method, skipping config save")
 
     if processor:
         LOG.info(f"Pre-saving processor to {cfg.output_dir}...")
@@ -484,12 +461,9 @@ def handle_untrained_tokens_fix(
     fix_untrained_tokens(model, tokenizer, train_dataset, **fix_kwargs)
 
     if cfg.local_rank == 0:
-        if hasattr(model, "save_pretrained"):
         model.save_pretrained(
             str(Path(cfg.output_dir)), safe_serialization=safe_serialization
         )
-        else:
-            LOG.warning("Model does not have save_pretrained method, skipping save")
 
 
 def setup_model_and_trainer(cfg: DictDefault, dataset_meta: TrainDatasetMeta) -> tuple[

src/axolotl/utils/callbacks/__init__.py

@@ -798,7 +798,7 @@ class SaveAxolotlConfigtoWandBCallback(TrainerCallback):
         control: TrainerControl,
         **kwargs,  # pylint: disable=unused-argument
     ):
-        if is_main_process():
+        if state.is_world_process_zero:
             try:
                 # sync config to top level in run, cannot delete file right away because wandb schedules it to be synced even w/policy = 'now', so let OS delete it later.
                 with NamedTemporaryFile(

src/axolotl/utils/config/__init__.py

@@ -148,8 +148,6 @@ def normalize_config(cfg):
                 f"Invalid value for eval_steps ({eval_steps}) from evals_per_epoch and/or num_epochs. Skipping evaluations."
             )
 
-    cfg.dataset_processes = cfg.dataset_processes or os.cpu_count()
-
     if not cfg.base_model_config:
         cfg.base_model_config = cfg.base_model
 

src/axolotl/utils/data/shared.py

@@ -410,9 +410,8 @@ def save_preprocessed_dataset(
 ) -> None:
     """Save preprocessed dataset to disk and optionally push to the HF Hub."""
     prepared_ds_path = get_prepared_dataset_path(cfg, dataset_hash)
-    if isinstance(dataset, IterableDataset):
     num_workers = cfg.dataset_processes
-
+    if isinstance(dataset, IterableDataset):
         ds_from_iter = Dataset.from_generator(
             functools.partial(_generate_from_iterable_dataset, dataset),
             features=dataset.features,
@@ -423,10 +422,20 @@ def save_preprocessed_dataset(
                 "num_workers": [num_workers] * num_workers,
             },
         )
-        ds_from_iter.save_to_disk(str(prepared_ds_path))
+        ds_from_iter.save_to_disk(
+            str(prepared_ds_path),
+            num_proc=num_workers,
+            max_shard_size=None,
+            num_shards=cfg.num_dataset_shards_to_save,
+        )
     else:
         os.makedirs(prepared_ds_path, exist_ok=True)
-        dataset.save_to_disk(str(prepared_ds_path))
+        dataset.save_to_disk(
+            str(prepared_ds_path),
+            num_proc=num_workers,
+            max_shard_size=None,
+            num_shards=cfg.num_dataset_shards_to_save,
+        )
     if cfg.push_dataset_to_hub:
         LOG.info(
             "Pushing merged prepared dataset to Huggingface hub at "
@@ -460,13 +469,13 @@ def load_preprocessed_dataset(cfg: DictDefault, dataset_hash: str) -> Dataset |
     ):
         LOG.info(
             f"Loading prepared dataset from disk at {prepared_ds_path}...",
-            main_process_only=False,
+            main_process_only=True,
         )
         return load_from_disk(str(prepared_ds_path))
 
     LOG.info(
         f"Unable to find prepared dataset in {prepared_ds_path}",
-        main_process_only=False,
+        main_process_only=True,
     )
     return None
 

src/axolotl/utils/quantization.py

@@ -20,6 +20,7 @@ from torchao.quantization.quant_api import (
     UIntXWeightOnlyConfig,
     _is_linear,
 )
+from transformers import TorchAoConfig
 
 from axolotl.utils.schemas.enums import TorchIntDType
 
@@ -149,7 +150,9 @@ def quantize_model_for_ptq(
         group_size=group_size,
     )
     quantize_(model, linear_ptq_config)
+    quantization_config = TorchAoConfig(linear_ptq_config)
     if quantize_embedding:
+        quantization_config.include_input_output_embeddings = True
         embedding_quantize_config = get_ptq_config(
             weight_dtype=weight_dtype,
             activation_dtype=None,
@@ -160,6 +163,7 @@ def quantize_model_for_ptq(
             embedding_quantize_config,
             filter_fn=lambda m, _: isinstance(m, torch.nn.Embedding),
         )
+        model.config.quantization_config = quantization_config
 
 
 def convert_qat_model_for_ptq(

src/axolotl/utils/schemas/config.py

@@ -193,6 +193,12 @@ class AxolotlInputConfig(
         json_schema_extra={"description": "Index of shard to use for whole dataset"},
     )
     skip_prepare_dataset: bool | None = False
+    num_dataset_shards_to_save: int | None = Field(
+        default=None,
+        json_schema_extra={
+            "description": "Number of shards to save the prepared dataset"
+        },
+    )
 
     pretraining_dataset: (
         Annotated[list[PretrainingDataset | SFTDataset], MinLen(1)] | None
@@ -203,11 +209,12 @@ class AxolotlInputConfig(
         },
     )
     dataset_processes: int | None = Field(
-        default=min(
+        default=None,
-            int(os.environ.get("AXOLOTL_DATASET_PROCESSES", 32)), os.cpu_count()
-        ),  # type: ignore[type-var]
         json_schema_extra={
-            "description": "The maximum number of processes to use while preprocessing your input dataset. This defaults to `os.cpu_count()` if not set."
+            "description": (
+                "The maximum number of processes to use while preprocessing your input dataset. This defaults to `os.cpu_count()` if not set.\n"
+                "For Runpod VMs, it will default to number of vCPUs via RUNPOD_CPU_COUNT."
+            )
         },
     )
     dataset_exact_deduplication: bool | None = Field(
@@ -1199,3 +1206,16 @@ class AxolotlConfigWCapabilities(AxolotlInputConfig):
             data["dataloader_prefetch_factor"] = 256
 
         return data
+
+    @model_validator(mode="before")
+    @classmethod
+    def default_dataset_processes(cls, data):
+        if data.get("dataset_processes") is None:
+            if axolotl_dataset_processes := os.environ.get("AXOLOTL_DATASET_PROCESSES"):
+                data["dataset_processes"] = int(axolotl_dataset_processes)
+            elif runpod_cpu_count := os.environ.get("RUNPOD_CPU_COUNT"):
+                data["dataset_processes"] = int(runpod_cpu_count)
+            else:
+                data["dataset_processes"] = os.cpu_count()
+
+        return data