ms-swift/docs/source_en/Instruction/Megatron-SWIFT-Training.md

Megatron-SWIFT Training

SWIFT incorporates Megatron's parallelization techniques to accelerate the training of large models, including data parallelism, tensor parallelism, pipeline parallelism, sequence parallelism, context parallelism, and expert parallelism. It supports the pre-training and fine-tuning of models such as Qwen3, Qwen3-MoE, Qwen2.5, Llama3, and the Deepseek-R1 series. For a complete list of supported models, please refer to the Supported Models and Datasets documentation.

Environment Setup

To use Megatron-SWIFT, in addition to installing the swift dependencies, you also need to install the following:

# Recommended PyTorch version: 2.5 / 2.6
pip install pybind11

# transformer_engine
# If an installation error occurs, you can refer to this issue for resolution: https://github.com/modelscope/ms-swift/issues/3793
pip install --no-build-isolation transformer_engine[pytorch]
# Or install using the following command
# pip install --no-build-isolation git+https://github.com/NVIDIA/TransformerEngine.git@release_v2.5#egg=transformer_engine[pytorch]

# apex
git clone https://github.com/NVIDIA/apex
cd apex
# https://github.com/modelscope/ms-swift/issues/4176
git checkout e13873debc4699d39c6861074b9a3b2a02327f92
pip install -v --disable-pip-version-check --no-cache-dir --no-build-isolation --config-settings "--build-option=--cpp_ext" --config-settings "--build-option=--cuda_ext" ./

# megatron-core
pip install git+https://github.com/NVIDIA/Megatron-LM.git@core_r0.13.0

# If you are using multi-node training, please additionally set the `MODELSCOPE_CACHE` environment variable to a shared storage path.
# This will ensure that the dataset cache is shared, thereby speeding up preprocessing.
export MODELSCOPE_CACHE='/xxx/shared'

Alternatively, you can also use the image:

modelscope-registry.cn-hangzhou.cr.aliyuncs.com/modelscope-repo/modelscope:ubuntu22.04-cuda12.4.0-py310-torch2.6.0-vllm0.8.5.post1-modelscope1.28.1-swift3.6.3
modelscope-registry.cn-beijing.cr.aliyuncs.com/modelscope-repo/modelscope:ubuntu22.04-cuda12.4.0-py310-torch2.6.0-vllm0.8.5.post1-modelscope1.28.1-swift3.6.3
modelscope-registry.us-west-1.cr.aliyuncs.com/modelscope-repo/modelscope:ubuntu22.04-cuda12.4.0-py310-torch2.6.0-vllm0.8.5.post1-modelscope1.28.1-swift3.6.3

The training module in the dependent library Megatron-LM will be cloned and installed by swift via git clone. Alternatively, you can use the environment variable MEGATRON_LM_PATH to point to the path of an already downloaded repository (in offline environments, use the core_r0.13.0 branch).

Quick Start Example

This section introduces a quick start example for fine-tuning the self-awareness of the Qwen2.5-7B-Instruct model using two 80GiB A100 GPUs. The following best practices can be completed within 10 minutes.

First, we need to convert the weights from HF (Hugging Face) format to Megatron format:

• If OOM (Out of Memory) occurs, simply remove CUDA_VISIBLE_DEVICES=0; the system will automatically use multiple GPUs. If you encounter insufficient memory, please remove --test_convert_precision true.

CUDA_VISIBLE_DEVICES=0 \
swift export \
    --model Qwen/Qwen2.5-7B-Instruct \
    --to_mcore true \
    --torch_dtype bfloat16 \
    --output_dir Qwen2.5-7B-Instruct-mcore \
    --test_convert_precision true

Next, use the following script to start training. The required GPU memory resources are 2*80GiB:

• If using multi-machine training, it is recommended to share a disk and specify the same path for --save.

PYTORCH_CUDA_ALLOC_CONF='expandable_segments:True' \
NPROC_PER_NODE=2 \
CUDA_VISIBLE_DEVICES=0,1 \
megatron sft \
    --load Qwen2.5-7B-Instruct-mcore \
    --dataset 'AI-ModelScope/alpaca-gpt4-data-zh#500' \
              'AI-ModelScope/alpaca-gpt4-data-en#500' \
              'swift/self-cognition#500' \
    --tensor_model_parallel_size 2 \
    --sequence_parallel true \
    --micro_batch_size 16 \
    --global_batch_size 16 \
    --recompute_granularity full \
    --recompute_method uniform \
    --recompute_num_layers 1 \
    --finetune true \
    --cross_entropy_loss_fusion true \
    --lr 1e-5 \
    --lr_warmup_fraction 0.05 \
    --min_lr 1e-6 \
    --max_epochs 1 \
    --save megatron_output/Qwen2.5-7B-Instruct \
    --save_interval 100 \
    --max_length 2048 \
    --system 'You are a helpful assistant.' \
    --num_workers 4 \
    --no_save_optim true \
    --no_save_rng true \
    --dataset_num_proc 4 \
    --model_author swift \
    --model_name swift-robot

Finally, convert the Megatron format weights back to HF format:

• Note: Please point --mcore_model to the parent directory of iter_xxx. By default, the corresponding checkpoint from latest_checkpointed_iteration.txt will be used.
• If OOM (Out of Memory) occurs, simply remove CUDA_VISIBLE_DEVICES=0. If you encounter insufficient memory, please remove --test_convert_precision true.

CUDA_VISIBLE_DEVICES=0 \
swift export \
    --mcore_model megatron_output/Qwen2.5-7B-Instruct/vx-xxx \
    --to_hf true \
    --torch_dtype bfloat16 \
    --output_dir megatron_output/Qwen2.5-7B-Instruct/vx-xxx-hf \
    --test_convert_precision true

We then perform inference on the generated HF format weights:

CUDA_VISIBLE_DEVICES=0 \
swift infer \
    --model megatron_output/Qwen2.5-7B-Instruct/vx-xxx-hf \
    --stream true \
    --temperature 0 \
    --max_new_tokens 2048

The inference results are as follows:

<<< who are you?
I am a language model developed by swift, you can call me swift-robot. How can I assist you?

• For pretraining, you can use megatron pt instead of megatron sft, which will use a generative template for training.
• More examples: Including packing, multi-node training, 32K context, DPO, MoE models, and pre-training, can be found here.
• The custom dataset format is the same as ms-swift. Refer to the custom dataset documentation.

LoRA Training

Best practice reference for single-node 8xH20 LoRA training with Qwen3-235B-A22B-Instruct-250718: https://github.com/modelscope/ms-swift/pull/5033.

Compared to full parameter tuning, LoRA training differs in both the training and MCore-to-HF conversion scripts:

Training Script:

# full: 2 * 70GiB 0.61s/it
# lora: 2 * 14GiB 0.45s/it
PYTORCH_CUDA_ALLOC_CONF='expandable_segments:True' \
NPROC_PER_NODE=2 \
CUDA_VISIBLE_DEVICES=0,1 \
megatron sft \
    --load Qwen2.5-7B-Instruct-mcore \
    --dataset 'AI-ModelScope/alpaca-gpt4-data-zh#500' \
              'AI-ModelScope/alpaca-gpt4-data-en#500' \
              'swift/self-cognition#500' \
    --train_type lora \
    --lora_rank 8 \
    --lora_alpha 32 \
    --target_modules all-linear \
    --tensor_model_parallel_size 2 \
    --sequence_parallel true \
    --micro_batch_size 16 \
    --global_batch_size 16 \
    --recompute_granularity full \
    --recompute_method uniform \
    --recompute_num_layers 1 \
    --finetune true \
    --cross_entropy_loss_fusion true \
    --lr 1e-4 \
    --lr_warmup_fraction 0.05 \
    --min_lr 1e-5 \
    --max_epochs 1 \
    --save megatron_output/Qwen2.5-7B-Instruct \
    --save_interval 100 \
    --max_length 2048 \
    --system 'You are a helpful assistant.' \
    --num_workers 4 \
    --no_save_optim true \
    --no_save_rng true \
    --dataset_num_proc 4 \
    --model_author swift \
    --model_name swift-robot

• For LoRA training scripts of MoE models, please refer to here.

MCore to HF Conversion Script:

CUDA_VISIBLE_DEVICES=0 \
swift export \
    --mcore_adapters megatron_output/Qwen2.5-7B-Instruct/vx-xxx \
    --to_hf true \
    --torch_dtype bfloat16 \
    --output_dir megatron_output/Qwen2.5-7B-Instruct/vx-xxx-hf \
    --test_convert_precision true

• Note: The mcore_adapters folder contains an args.json file. During the conversion process, parameters related to mcore_model and LoRA will be loaded from this file. The system will then perform a merge-lora operation between the mcore_model and mcore_adapters to obtain the complete model weights, and finally convert them into HuggingFace (HF) format.

Benchmark

The speed comparison of full-parameter training for Dense/MoE models using megatron sft and swift sft on a single machine with eight A800 GPUs is shown below. The corresponding scripts can be found here.

Dense Qwen2.5-14B:

	Megatron-LM	Deepspeed-ZeRO2	Deepspeed-ZeRO3
Training Speed	9.04s/it	10.32s/it	10.56s/it
GPU Memory Usage	8*64GB	8*80GB	8*58GB

MoE Qwen1.5-MoE-A2.7B:

	Megatron-LM	Deepspeed-ZeRO2	Deepspeed-ZeRO3
Training Speed	2.95s/it	6.02s/it	24.30s/it
GPU Memory Usage	8*57GB	8*72GB	8*50GB

Command Line Arguments

Megatron Parameters

Training Parameters:

• 🔥micro_batch_size: Batch size per device, default is 1.
• 🔥global_batch_size: Total batch size, equivalent to micro_batch_size * data parallel size * gradient accumulation steps. Default is 16.
• 🔥recompute_granularity: Granularity of activation recomputation, options are 'full', 'selective'. 'full' means recomputing the entire transformer layer, while 'selective' means only recomputing the core attention part of the transformer layer. 'selective' is generally recommended. Default is 'selective'.
• 🔥recompute_method: This parameter takes effect only when recompute_granularity is set to 'full', options are 'uniform', 'block'. Default is None.
• 🔥recompute_num_layers: This parameter takes effect only when recompute_granularity is set to 'full'. Default is None. If recompute_method is set to uniform, this parameter specifies the number of transformer layers in each uniformly divided recomputation unit. For example, you can specify --recompute_granularity full --recompute_method uniform --recompute_num_layers 4. The larger the recompute_num_layers, the smaller the memory usage but higher computation cost. Default is None.
• recompute_modules: Options include "core_attn", "moe_act", "layernorm", "mla_up_proj", "mlp", and "moe". The default value is ["core_attn"]. This parameter takes effect when --recompute_granularity selective is set. For example, during MoE training, you can reduce memory usage by specifying --recompute_granularity selective --recompute_modules core_attn moe. Among these, "core_attn", "mlp", and "moe" use normal checkpointing, while "moe_act", "layernorm", and "mla_up_proj" use output-discarding checkpointing.
  • "core_attn": Recomputes the core attention part of the Transformer layer.
  • "mlp": Recomputes the dense MLP layer.
  • "moe": Recomputes the MoE layer.
  • "moe_act": Recomputes the MLP activation function part in the MoE module.
  • "layernorm": Recomputes the input_layernorm and pre_mlp_layernorm.
  • "mla_up_proj": Recomputes the MLA up-projection and RoPE application parts.
• deterministic_mode: Deterministic mode, which may lead to slower training speed, default is False.
• 🔥train_iters: Total number of training iterations, default is None.
• 🔥log_interval: Log interval (unit: iters), default is 5.
• tensorboard_dir: Directory where TensorBoard logs are written. Default is None, meaning logs will be stored in the f'{save}/runs' directory.
• no_masked_softmax_fusion: Default is False. Disables scaling, masking, and softmax fusion for query_key_value.
• no_bias_dropout_fusion: Default is False. Disables bias and dropout fusion.
• no_bias_swiglu_fusion: Default is False. Specify --no_bias_dropout_fusion true to disable bias and swiglu fusion.
• no_rope_fusion: Default is False. Specify --no_rope_fusion true to disable rope fusion.
• no_gradient_accumulation_fusion: Default is False. Specify --no_gradient_accumulation_fusion true to disable gradient accumulation fusion.
• 🔥cross_entropy_loss_fusion: Enables cross-entropy loss calculation fusion. Default is False.
• cross_entropy_fusion_impl: Implementation of cross-entropy loss fusion. Options include 'native' and 'te'. Defaults to 'native'.
• calculate_per_token_loss: Scales the cross-entropy loss according to the number of non-padded tokens in the global batch. Default is True.
  • Note: The default is False in RLHF.
• 🔥attention_backend: The attention backend to use (flash, fused, unfused, local, auto). Defaults to auto.
• optimizer: Optimizer type, options are 'adam', 'sgd'. Default is adam.
• 🔥optimizer_cpu_offload: Offloads the optimizer state to CPU. Default is False.
• optimizer_offload_fraction: The fraction of the optimizer state to offload to CPU. Default is 1.0.
• use_precision_aware_optimizer: Use the precision-aware optimizer in TransformerEngine, which allows setting the main parameters and optimizer states to lower precision, such as fp16 and fp8.
• main_grads_dtype: The dtype of main gradients when use_precision_aware_optimizer is enabled. Options are 'fp32' and 'bf16'. Default is 'fp32'.
• main_params_dtype: The dtype of main parameters when use_precision_aware_optimizer is enabled. Options are 'fp32' and 'fp16'. Default is 'fp32'.
• exp_avg_dtype: The dtype of exp_avg (i.e., the first moment in the Adam optimizer) when use_precision_aware_optimizer is enabled. This dtype is used for storing the optimizer state in memory during training, but does not affect the precision in kernel computation. Options are 'fp32', 'fp16', 'bf16', and 'fp8'. Default is 'fp32'.
• exp_avg_sq_dtype: The dtype of exp_avg_sq (i.e., the second moment in the Adam optimizer) when use_precision_aware_optimizer is enabled. This dtype is used for storing the optimizer state in memory during training, but does not affect the precision in kernel computation. Options are 'fp32', 'fp16', 'bf16', and 'fp8'. Default is 'fp32'.
• dataloader_type: Default is 'cyclic', options are 'single', 'cyclic', 'external'. If --streaming is enabled, set it to external.
• manual_gc: Disables the default garbage collector and manually triggers garbage collection. Default is False.
• manual_gc_interval: Interval at which garbage collection is triggered. Default is 0.
• seed: Random seed for python, numpy, pytorch, and cuda, default is 42.
• 🔥num_workers: Number of workers for the dataloader, default is 4.
  • Note: If --streaming true is set, it will be set to 1. seq_length: Defaults to None, meaning it is set to max_length. To restrict the dataset length, please use the --max_length parameter in the basic arguments; there is no need to set this parameter.
• use_cpu_initialization: Initializes weights on the CPU, default is False. Used during HF and MCore weight conversion.
• extra_megatron_kwargs: Additional parameters passed to Megatron, provided as a JSON object. Defaults to None.

Learning Rate Parameters:

• 🔥lr: The initial learning rate. The actual learning rate for each iteration will be determined based on the learning rate warmup and decay strategies. The default value is None; for full-parameter training, the default is 1e-5, while for LoRA training, the default is 1e-4.
• lr_decay_style: Learning rate decay strategy, default is 'cosine'. Commonly set to 'cosine', 'linear', or 'constant'.
• 🔥lr_decay_iters: Number of iterations for learning rate decay. Default is None, meaning it will be set to --train_iters.
• lr_warmup_iters: Number of iterations for linear learning rate warm-up, default is 0.
• 🔥lr_warmup_fraction: The fraction of the linear learning rate warmup phase, defaults to None.
• 🔥min_lr: Minimum value of the learning rate, clipping any learning rate below this threshold to this value, default is 0.

Regularization Parameters:

• 🔥weight_decay: Default is 0.1.
• 🔥clip_grad: L2 gradient clipping, default is 1.0.
• adam_beta1: Default is 0.9.
• adam_beta2: Default is 0.95.
• adam_eps: Default is 1e-8.
• sgd_momentum: Default is 0.9.

Checkpoint Parameters:

• 🔥save: Output directory for checkpoints, default is None. During training, if this parameter is not set, it defaults to f'megatron_output/{model_suffix}', e.g., 'megatron_output/Qwen2.5-7B-Instruct'.
  • Note: When training on multiple machines, ensure that the save paths on each node point to the same location. Otherwise, you will need to manually consolidate these weights after training.
• 🔥save_interval: Checkpoint saving interval (steps), default is 500.
  • Note: Weights will always be saved at the end of training.
• 🔥no_save_optim: Do not save optimizer, default is False.
• 🔥no_save_rng: Do not save RNG, default is False.
• 🔥load: Directory of the checkpoint to load, default is None.
• 🔥no_load_optim: Do not load optimizer, default is False.
• 🔥no_load_rng: Do not load RNG, default is False.
• 🔥finetune: Load and fine-tune the model. Optimizer and random seed states from the checkpoint will not be loaded, and the number of iterations will be set to 0. The default is False.
  • Note: For checkpoint resumption (--load), if --finetune true is set, the dataset will not be skipped; if not set, previously trained datasets will be skipped.
  • Streaming datasets (--streaming) are currently not supported for skipping datasets.
• ckpt_format: Format of the checkpoint. Options are 'torch', 'torch_dist', 'zarr'. Default is 'torch_dist'.
• no_initialization: Do not initialize weights, default is True.
• auto_detect_ckpt_format: Automatically detect whether the checkpoint format is legacy or distributed. Default is True.
• exit_on_missing_checkpoint: If --load is set but no checkpoint is found, exit directly instead of initializing. Default is True.

Distributed Parameters:

• distributed_backend: Distributed backend, options are 'nccl', 'gloo'. Default is nccl.
• 🔥use_distributed_optimizer: Use a distributed optimizer. Default is True.
• 🔥tensor_model_parallel_size: TP (Tensor Parallelism) size, default is 1.
• 🔥pipeline_model_parallel_size: PP (Pipeline Parallelism) size, default is 1.
• 🔥decoder_first_pipeline_num_layers: The number of Transformer layers in the first pipeline stage of the decoder. Default is None, which means the Transformer layers are evenly distributed across all pipeline stages.
• 🔥decoder_last_pipeline_num_layers: The number of Transformer layers in the last pipeline stage of the decoder. Default is None, which means the Transformer layers are evenly distributed across all pipeline stages.
• 🔥sequence_parallel: Enable sequence parallel optimization. Default is False.
• 🔥context_parallel_size: CP (Context Parallelism) size, default is 1.
• tp_comm_overlap: Overlap tensor parallel communication with GEMM (General Matrix Multiplication) kernels (to reduce communication time). Default is False.
• 🔥overlap_grad_reduce: Overlap grad reduction operations in DDP (to reduce DP communication time). Default is False.
• 🔥overlap_param_gather: Overlap all-gather of parameters in the distributed optimizer (to reduce DP communication time). Default is False.
• distributed_timeout_minutes: The timeout duration for torch.distributed (in minutes). This parameter is deprecated and is now controlled by the ddp_timeout in the Base Arguments, with a default value of 300000 minutes.

Logging Parameters:

• log_params_norm: Logs the norm of parameters. Default is False.
• log_throughput: Logs throughput per GPU. Default is False.
  • Note: In non-packing scenarios, log_throughput is not accurate because seq_length does not equal the actual sequence length.
• tensorboard_log_interval: Interval (steps) for logging to TensorBoard, default is 1.
• tensorboard_queue_size: Queue length (related to disk I/O), similar to write intervals. Default is 50.
• log_timers_to_tensorboard: Logs timers to TensorBoard. Default is True.
• no_log_learning_rate_to_tensorboard: Do not log learning rate to TensorBoard. Default is False.
• log_validation_ppl_to_tensorboard: Writes validation perplexity to TensorBoard. Default is True.
• log_memory_to_tensorboard: Writes memory logs to TensorBoard. Default is True.
• logging_level: Logging level. Default is None.
• wandb_project: The name of the wandb project. Defaults to '', which means ignoring wandb.
• wandb_exp_name: The name of the wandb experiment. Defaults to ''.
• wandb_save_dir: The local path to save wandb results. Defaults to ''.

Evaluation Parameters:

• 🔥eval_iters: The number of iterations for evaluation. Defaults to -1, and a suitable value will be set based on the size of the validation dataset.
  • Note: If using a streaming dataset, this value needs to be set manually.
• 🔥eval_interval: The evaluation interval (steps), i.e., how many steps between each evaluation. The default is None, which means it will be set to save_interval.

FP8 Parameters:

• fp8_format: The FP8 format scheme used for FP8 tensors in the forward and backward pass. Options are 'e4m3' and 'hybrid'. Default is None.
• fp8_recipe: The FP8 recipe (algorithm scheme) used for FP8 tensors in the forward and backward pass. Options are 'tensorwise', 'delayed', 'mxfp8', and 'blockwise'. Default is 'delayed'.
• fp8_amax_history_len: Number of steps for which amax history is recorded per tensor. Default is 1024.
• fp8_amax_compute_algo: Algorithm for computing amax from history. Options are 'most_recent' and 'max'. Default is 'max'.
• fp8_param_gather: Keep the compute parameter in FP8 (do not use any other intermediate dtype) and perform the parameter all-gather in FP8 format. Default is False.

Mixed Precision Parameters:

• fp16: FP16 mode. The default is None, and it will be set according to the model's torch_dtype. The torch_dtype is read from the config.json by default.
• bf16: BF16 mode. The default is None, and it will be set according to the model's torch_dtype.
• apply_query_key_layer_scaling: Scales Q * K^T by 1 / layer number (e.g., divide by layer_num for layer_num-th layer). This is helpful for FP16 training. Default is None, meaning that if --fp16 is used, it will be set to True.
• attention_softmax_in_fp32: Uses FP32 for computations in attention_mask and softmax. Default is True.

Model Parameters: (The following parameters typically do not need to be set as they will be configured based on the HF model’s config.json; users don’t need to worry about them)

• num_layers: Number of transformer layers, default is None.
• hidden_size: Transformer hidden size, default is None.
• ffn_hidden_size: Hidden size of the FFN layer in the transformer. Default is None, set to 4*hidden_size.
• num_attention_heads: Number of transformer attention heads, default is None.
• group_query_attention: Default is None. If num_query_groups > 1, group_query_attention is set to True, otherwise False.
• num_query_groups: Default is 1.
• max_position_embeddings: Maximum length of positional embeddings, default is None.
• position_embedding_type: Type of positional embedding, options are 'learned_absolute', 'rope', 'mrope', 'relative', and 'none'. Default is 'rope'.
• rotary_base: Default is 10000.
• rotary_percent: Default is 1.
• normalization: Options are 'LayerNorm', 'RMSNorm'. Default is RMSNorm.
• norm_epsilon: Default is 1e-5.
• swiglu: Uses swiglu instead of the default gelu. Default is True.
• untie_embeddings_and_output_weights: Unties embedding and output weights. Default is True.
• disable_bias_linear: Disables bias in linear layers. Default is True.
• add_qkv_bias: Adds bias only to QKV linear layers. Default is True.
• attention_dropout: Default is 0.
• hidden_dropout: Default is 0.
• kv_channels: Defaults to None, set to args.hidden_size // args.num_attention_heads.
• qk_layernorm: Whether to apply layer normalization to Q and K.
• transformer_impl: Which transformer implementation to use, options are 'local' and 'transformer_engine'. Default is transformer_engine.
• padded_vocab_size: Full vocabulary size, default is None.
• rope_scaling: Related parameters for rope_scaling, default is None. Refer to the format in llama3.1 config.json. Pass the value as a JSON string.

MoE Parameters:

• num_experts: The number of experts in MoE, default is None. Automatically read from config.json.
• moe_layer_freq: Frequency distribution between MoE layers and Dense layers. Default is None. This parameter is read from config.json.
• moe_ffn_hidden_size: Hidden layer size of the feedforward network (ffn) for each expert. Default is None and will be automatically read from config.json. If not found and num_experts is not None, it will be set to ffn_hidden_size.
• moe_shared_expert_intermediate_size: The total FFN hidden layer size for shared experts. If there are multiple shared experts, it should equal num_shared_experts * ffn_size_of_each_shared_expert. Default is None. Automatically read from config.json.
• moe_router_topk: The number of experts each token is routed to. Default is None. Automatically read from config.json.
• moe_router_pre_softmax: Enable pre-softmax routing for MoE, meaning that softmax will be applied before top-k selection. Default is None. Automatically read from config.json.
• 🔥moe_router_dtype: Data type used for routing computation and expert output weighted averaging. Options are 'none', 'fp32', and 'fp64', which enhances numerical stability, especially when the number of experts is large. When used together with moe_permute_fusion, the performance impact is negligible. Default is 'fp32'. 'none' means no change to data type.
• moe_router_score_function: Scoring function for MoE TopK routing. Can be "softmax" or "sigmoid". Default is None and is read from config.json.
• moe_router_bias_update_rate: Update rate of expert bias in the auxiliary-loss-free load balancing strategy. Expert bias is updated based on the number of tokens each expert is assigned in the global batch: bias increases for experts assigned fewer tokens, and decreases for those assigned more tokens. Default is 1e-3, same as used in DeepSeekV3.
• moe_router_enable_expert_bias: TopK routing with dynamic expert bias in the auxiliary-loss-free load balancing strategy. Routing decisions are based on the sum of routing scores and expert bias. See details at: https://arxiv.org/abs/2408.15664. Default is None and is automatically read from config.json.
• moe_router_topk_scaling_factor: Default is None. This parameter is read from config.json.
• moe_router_load_balancing_type: Determines the router’s load balancing strategy. Options are "aux_loss", "seq_aux_loss", "sinkhorn", and "none". Default is None and is read from config.json.
• 🔥expert_model_parallel_size: The degree of expert parallelism, default is 1.
• moe_token_dispatcher_type: The type of token dispatcher to use. Options include 'allgather', 'alltoall', 'flex', and 'alltoall_seq'. Default is 'alltoall'.
• moe_enable_deepep: Experimental feature, Enables DeepSeek/DeepEP for efficient token dispatching and combination in MoE models. Only works when using the flexible token dispatcher by setting --moe_token_dispatcher_type flex.
• 🔥moe_grouped_gemm: When each rank contains multiple experts, multiple local GEMM kernels can be launched in parallel streams to improve utilization and performance by using GroupedLinear from TransformerEngine. Default is False.
• 🔥moe_permute_fusion: Fuses token permutation operations during token dispatch. Default is False.
• 🔥moe_aux_loss_coeff: Scaling coefficient for the auxiliary loss; a recommended initial value is 1e-2. Default is None and is automatically read from config.json.
• moe_z_loss_coeff: Scaling coefficient for z-loss. Default is None.
• moe_expert_capacity_factor: Capacity factor for each expert. None means no token will be dropped. Default is None and will be automatically read from config.json.
• 🔥moe_shared_expert_overlap: Enables overlap between shared expert computation and the dispatcher. If not enabled, shared expert computation will be performed after routing experts. Only effective when moe_shared_expert_intermediate_size is set. Default is False.
• moe_token_drop_policy: Options are 'probs' and 'position'. Default is 'probs'.

MLA Parameters

• multi_latent_attention: Whether to use MLA. Default is False.
• q_lora_rank: Low-rank representation rank value of the Query tensor. Default is None and will be automatically read from config.json.
• kv_lora_rank: Low-rank representation rank value of the Key and Value tensors. Default is None and will be automatically read from config.json.
• qk_head_dim: Dimension of the head in the QK projection. q_head_dim = qk_head_dim + qk_pos_emb_head_dim. Default is None and will be automatically read from config.json.
• qk_pos_emb_head_dim: Dimension of the position embedding in the QK projection. Default is None and will be automatically read from config.json.

Tuner Parameters:

• train_type: Options are 'lora' and 'full'. Default is 'full'.

Full-parameter Training:

• freeze_parameters: Prefixes of parameters to be frozen. Default is [].
• freeze_parameters_regex: Regex expression for parameters to be frozen. Default is None.
• freeze_parameters_ratio: The proportion of parameters to freeze from bottom to top. Default is 0. Setting this to 1 will freeze all parameters; you can set trainable parameters separately using trainable_parameters. This parameter is incompatible with PP (pipeline parallel) mode.
• trainable_parameters: Prefixes of additional trainable parameters. Default is [].
• trainable_parameters_regex: Regex expression to match additional trainable parameters. Default is None.

LoRA Training:

• adapter_load: The path to the adapter weights for loading, used for resuming LoRA training from a checkpoint. The default is None. The method for resuming LoRA training from a checkpoint is the same as for full-parameter training. Please pay attention to the meaning of the --finetune parameter.
• 🔥target_modules: Suffixes of modules to apply LoRA to. Default is ['all-linear'].
• 🔥target_regex: Regex expression to specify LoRA modules. Default is None. If this value is provided, the target_modules parameter will be ignored.
• 🔥modules_to_save: After attaching a tuner, explicitly specifies additional original model modules to participate in training and storage. The default is [].
• 🔥lora_rank: Default is 8.
• 🔥lora_alpha: Default is 32.
• lora_dropout: Default is 0.05.
• lora_bias: Default is 'none'. Available options: 'none', 'all'. If you want all biases to be set as trainable, set this to 'all'.
• use_rslora: Default is False. Whether to use RS-LoRA.

DPO Parameters

• ref_load: The path to load the reference model. Defaults to None, which means it will be set to load.
• beta: Has the same meaning as in TRL. It controls the degree of deviation from the reference model. A higher beta value indicates less deviation from the reference model. For the IPO loss function (loss_type="ipo"), beta is the regularization parameter as mentioned in the paper. Default is 0.1.
• rpo_alpha: A parameter from the RPO paper used to control the weight of the NLL term (i.e., SFT loss) in the loss function. The total loss is calculated as loss = dpo_loss + rpo_alpha * nll_loss. Default is 1.
• reference_free: Whether to ignore the provided reference model and implicitly use a reference model that assigns equal probability to all responses. Default is False.
• label_smoothing: Default is 0.
• f_divergence_type: Default is reverse_kl. See the TRL documentation for possible values.
• loss_type: Default is 'sigmoid'. See the TRL documentation for possible values.

Training Parameters

Megatron training parameters inherit from Megatron parameters and basic parameters. For information on basic parameters, see here. Additionally, the following parameters are included:

• add_version: Adds a directory <version>-<timestamp> to save to prevent overwriting weights, default is True.
• padding_free: Flattens the data in a batch to avoid padding, thereby reducing memory usage and accelerating training. Default is True.
  • If you wish to customize the attention_mask, you can set --padding_free false.
• mlp_padding_free: The default is False. This is used for applying padding-free optimization to the MLP when padding_free is set to false. It allows for improved training speed and reduced memory usage while customizing the attention_mask.
• 🔥packing: Whether to use sequence packing, defaults to False. Currently supports megatron pt/sft.
• packing_cache: Specifies the directory for packing cache. The default value is None, which means the cache will be stored in the path defined by the environment variable $MODELSCOPE_CACHE. When using the packing feature across multiple nodes, ensure that all nodes share the same packing cache directory. You can achieve this by setting the MODELSCOPE_CACHE environment variable or by adding the --packing_cache <shared_path> argument in the command line.
  • Note: This parameter will be removed in "ms-swift>=3.7". The packing_cache setting will no longer be required for multi-node packing.
• 🔥streaming: Stream reading and processing of the dataset, default is False. It is typically set to True when handling large datasets. For more information on streaming parameters, refer to the command-line parameters documentation.
• lazy_tokenize: Default is False. If this parameter is set to False, all dataset samples are tokenized before training (this avoids errors during training); if set to True, tokenization occurs during training (this saves memory).
• max_epochs: Forces the training to exit after reaching max_epochs, and performs validation and saving of the model weights. This parameter is especially useful when using a streaming dataset. Default is None.
  • Note: If you use a non-streaming dataset, this parameter will automatically calculate train_iters for you, so there is no need to pass train_iters manually.

RLHF Parameters

In addition to inheriting the training parameters, the following parameters are also supported:

• rlhf_type: Default is 'dpo'. Currently, only 'dpo' is available.
• loss_scale: Overrides the loss_scale in basic parameters. Default is 'last_round'.
• calculate_per_token_loss: Overrides the Megatron parameter. Default is False.