| import os |
|
|
| import pytest |
| import torch |
| import torch.nn as nn |
| import torch.nn.functional as F |
| from torch.optim import SGD, Adam |
|
|
| |
| from transformer_engine.pytorch.fp8 import fp8_autocast |
|
|
| from megatron.core.distributed import DistributedDataParallel, DistributedDataParallelConfig |
| from megatron.core.optimizer import ChainedOptimizer, OptimizerConfig, get_megatron_optimizer |
| from megatron.core.process_groups_config import ProcessGroupCollection |
| from megatron.core.transformer import TransformerConfig |
| from megatron.core.utils import is_te_min_version, is_torch_min_version |
| from tests.unit_tests.test_utilities import Utils |
| from tests.unit_tests.test_utils import _deinit_distributed, _init_distributed |
|
|
| try: |
| |
| from transformer_engine.pytorch.fp8 import check_fp8_block_scaling_support |
|
|
| fp8_block_scaling_available, reason_for_no_fp8_block_scaling = check_fp8_block_scaling_support() |
| from transformer_engine.common.recipe import Float8BlockScaling, Format |
| except: |
| fp8_block_scaling_available = False |
| reason_for_no_fp8_block_scaling = "FP8 block scaled GEMM requires Hopper and CUDA >= 12.9." |
| try: |
| from transformer_engine.common.recipe import DelayedScaling |
| except: |
| delayed_scaling_available = False |
|
|
|
|
| class Net(nn.Module): |
| def __init__(self): |
| super().__init__() |
| self.conv1 = nn.Conv2d(3, 6, 5) |
| self.pool = nn.MaxPool2d(2, 2) |
| self.conv2 = nn.Conv2d(6, 16, 5) |
| self.fc1 = nn.Linear(16 * 5 * 5, 120) |
| self.fc2 = nn.Linear(120, 84) |
| self.fc3 = nn.Linear(84, 10) |
|
|
| def forward(self, x): |
| x = self.pool(F.relu(self.conv1(x))) |
| x = self.pool(F.relu(self.conv2(x))) |
| x = torch.flatten(x, 1) |
| x = F.relu(self.fc1(x)) |
| x = F.relu(self.fc2(x)) |
| x = self.fc3(x) |
| return x |
|
|
|
|
| def test_chained_optimizer(): |
| net = Net() |
| optimizer_1 = Adam(list(net.parameters())[:2], lr=0.01) |
| optimizer_2 = SGD(list(net.parameters())[2:], lr=0.1, momentum=0.9) |
| chained_optimizer = ChainedOptimizer([optimizer_1, optimizer_2]) |
|
|
| |
| assert optimizer_1.param_groups[0]["lr"] == 0.01 |
| chained_optimizer.param_groups[0]["lr"] = 0.02 |
| assert optimizer_1.param_groups[0]["lr"] == 0.02 |
|
|
| |
| |
| assert len(chained_optimizer.state) == 0 |
| input = torch.randn(1, 3, 32, 32) |
| output = net(input) |
| output.sum().backward() |
| optimizer_1.step() |
| optimizer_2.step() |
| assert len(chained_optimizer.state) != 0 |
|
|
| |
| assert not list(optimizer_1.state.values())[0]["exp_avg"].is_cuda |
| assert not list(optimizer_2.state.values())[0]["momentum_buffer"].is_cuda |
|
|
| def to_cuda(d): |
| for k, v in d.items(): |
| if isinstance(v, torch.Tensor): |
| d[k] = v.to("cuda") |
| elif isinstance(v, dict): |
| to_cuda(v) |
| return d |
|
|
| for k, v in chained_optimizer.state.items(): |
| chained_optimizer.state[k] = to_cuda(v) |
|
|
| assert list(optimizer_1.state.values())[0]["exp_avg"].is_cuda |
| assert list(optimizer_2.state.values())[0]["momentum_buffer"].is_cuda |
|
|
|
|
| def test_precision_aware_fused_adam(): |
| try: |
| from transformer_engine.pytorch.optimizers import FusedAdam |
| except ImportError: |
| |
| return |
|
|
| import inspect |
|
|
| adam_args = inspect.signature(FusedAdam).parameters |
| arg_names = ["master_weight_dtype", "exp_avg_dtype", "exp_avg_sq_dtype", "use_decoupled_grad"] |
| for name in arg_names: |
| if name not in adam_args: |
| |
| return |
|
|
| tensor = torch.rand(278011, dtype=torch.bfloat16).cuda() |
| params_1 = [torch.nn.Parameter(tensor.float())] |
| params_2 = [torch.nn.Parameter(tensor.clone())] |
|
|
| options = {"lr": 1, "betas": (0.1, 0.25), "eps": 1e-08, "weight_decay": 0, "amsgrad": False} |
|
|
| optimizer_1 = FusedAdam(params_1, **options) |
| optimizer_2 = FusedAdam(params_2, master_weights=True, use_decoupled_grad=True, **options) |
|
|
| for _ in range(1000): |
| for p_1, p_2 in zip(params_1, params_2): |
| p_1.grad = torch.rand_like(p_1) |
| p_2.decoupled_grad = p_1.grad.clone() |
|
|
| optimizer_1.step() |
| optimizer_2.step() |
|
|
| master_params = [optimizer_2.get_unscaled_state(p, "master_param") for p in params_2] |
| for p_1, p_2 in zip(params_1, master_params): |
| bytes_1 = p_1.data.view(torch.uint8) |
| bytes_2 = p_2.data.view(torch.uint8) |
| |
| assert torch.all(bytes_1 == bytes_2) |
|
|
| for p_1, p_2 in zip(params_1, params_2): |
| bytes_1 = p_1.data.bfloat16().view(torch.uint8) |
| bytes_2 = p_2.data.view(torch.uint8) |
| |
| assert torch.all(bytes_1 == bytes_2) |
|
|
|
|
| @pytest.mark.skipif( |
| not is_te_min_version("1.13.0"), reason="TE 1.13.0 is required for precision aware optimizer" |
| ) |
| @pytest.mark.parametrize("precision", ['bf16', 'fp8']) |
| @pytest.mark.parametrize("main_params_dtype", [torch.float32, torch.float16]) |
| @pytest.mark.parametrize("main_grads_dtype", [torch.float32, torch.bfloat16]) |
| @pytest.mark.parametrize( |
| |
| "moment_dtype", |
| [torch.float32, torch.float16, torch.bfloat16, torch.uint8], |
| ) |
| def test_precision_aware_optimizer( |
| precision: str, |
| main_params_dtype: torch.dtype, |
| main_grads_dtype: torch.dtype, |
| moment_dtype: torch.dtype, |
| ): |
| |
| if (moment_dtype == torch.bfloat16) and not is_te_min_version("2.3.0"): |
| pytest.skip("bfloat16 for moment_dtype requires TE >= 2.3.0") |
|
|
| if precision == 'fp8': |
| if not fp8_block_scaling_available: |
| fp8_recipe = "delayed" |
| fp8_recipe_settings = DelayedScaling() |
| else: |
| fp8_recipe = "blockwise" |
| fp8_recipe_settings = Float8BlockScaling(fp8_format=Format.E4M3) |
| else: |
| fp8_recipe = None |
| fp8_recipe_settings = None |
|
|
| world = int(os.getenv('WORLD_SIZE', '1')) |
| rank = int(os.getenv('RANK', '0')) |
|
|
| |
| _init_distributed(world, rank) |
| Utils.initialize_model_parallel() |
|
|
| |
| baseline_model = torch.nn.Linear(100, 100, bias=False, dtype=torch.bfloat16, device='cuda') |
| baseline_model.requires_grad_(True) |
| baseline_model.weight.data.fill_(1.0) |
| baseline_ddp_config = DistributedDataParallelConfig(use_distributed_optimizer=True) |
| baseline_model = DistributedDataParallel( |
| TransformerConfig(num_attention_heads=1, num_layers=1), baseline_ddp_config, baseline_model |
| ) |
| baseline_optimizer_config = OptimizerConfig( |
| optimizer='adam', |
| lr=0.01, |
| bf16=True, |
| use_distributed_optimizer=True, |
| use_precision_aware_optimizer=False, |
| main_params_dtype=torch.float32, |
| main_grads_dtype=torch.float32, |
| exp_avg_dtype=torch.float32, |
| exp_avg_sq_dtype=torch.float32, |
| ) |
| baseline_optim = get_megatron_optimizer(baseline_optimizer_config, [baseline_model]) |
|
|
| |
| test_model = torch.nn.Linear(100, 100, bias=False, dtype=torch.bfloat16, device='cuda') |
| test_model.requires_grad_(True) |
| test_model.weight.data.fill_(1.0) |
| ddp_config = DistributedDataParallelConfig(use_distributed_optimizer=True) |
| test_model = DistributedDataParallel( |
| TransformerConfig(num_attention_heads=1, num_layers=1), ddp_config, test_model |
| ) |
| test_optimizer_config = OptimizerConfig( |
| optimizer='adam', |
| lr=0.01, |
| bf16=True, |
| fp8_recipe=fp8_recipe, |
| use_distributed_optimizer=True, |
| use_precision_aware_optimizer=True, |
| main_params_dtype=main_params_dtype, |
| main_grads_dtype=main_grads_dtype, |
| exp_avg_dtype=moment_dtype, |
| exp_avg_sq_dtype=moment_dtype, |
| ) |
| test_optim = get_megatron_optimizer(test_optimizer_config, [test_model]) |
|
|
| |
| input = torch.randn(8, 100, dtype=torch.bfloat16, device='cuda') |
|
|
| |
| def run_model(model, input, optim, fp8_recipe, fp8_recipe_settings): |
| if not fp8_recipe: |
| output = model(input) |
| else: |
| with fp8_autocast(enabled=True, fp8_recipe=fp8_recipe_settings): |
| output = model(input) |
| loss = output.sum() |
| loss.backward() |
| optim.step() |
| return loss.item(), optim.get_grad_norm() |
|
|
| |
| baseline_loss, baseline_grad_norm = run_model( |
| baseline_model, input, baseline_optim, fp8_recipe, fp8_recipe_settings |
| ) |
| test_loss, test_grad_norm = run_model( |
| test_model, input, test_optim, fp8_recipe, fp8_recipe_settings |
| ) |
|
|
| rtol = 1e-3 |
| atol = 1e-5 |
|
|
| |
| rel_diff = abs(test_grad_norm - baseline_grad_norm) / ( |
| abs(baseline_grad_norm) + 1e-7 |
| ) |
| abs_diff = abs(test_grad_norm - baseline_grad_norm) |
| assert ( |
| rel_diff <= rtol or abs_diff <= atol |
| ), f"Grad norm mismatch: baseline={baseline_grad_norm}, test={test_grad_norm}, rel_diff={rel_diff}, abs_diff={abs_diff}" |
|
|
| |
| loss_rel_diff = abs(test_loss - baseline_loss) / (abs(baseline_loss) + 1e-7) |
| loss_abs_diff = abs(test_loss - baseline_loss) |
| assert ( |
| loss_rel_diff <= rtol or loss_abs_diff <= atol |
| ), f"Loss mismatch: baseline={baseline_loss}, test={test_loss}, rel_diff={loss_rel_diff}, abs_diff={loss_abs_diff}" |
|
|
| |
| state_dict = test_optim.state_dict() |
| test_optim.load_state_dict(state_dict) |
|
|
|
|
| @pytest.mark.parametrize("use_distributed_optimizer", [False, True]) |
| @pytest.mark.parametrize("precision", ['bf16', 'fp32']) |
| def test_optim_sharded_state_dict(use_distributed_optimizer: bool, precision: str): |
| world = int(os.getenv('WORLD_SIZE', '1')) |
| rank = int(os.getenv('RANK', '0')) |
|
|
| |
| _init_distributed(world, rank) |
| Utils.initialize_model_parallel() |
| model = torch.nn.Linear(100, 100, bias=False, dtype=torch.bfloat16, device='cuda') |
| model.requires_grad_(True) |
| model.weight.data.fill_(1.0) |
| ddp_config = DistributedDataParallelConfig(use_distributed_optimizer=use_distributed_optimizer) |
| model = DistributedDataParallel( |
| TransformerConfig(num_attention_heads=1, num_layers=1), ddp_config, model |
| ) |
| for param in model.parameters(): |
| assert param.requires_grad |
|
|
| if precision == 'bf16': |
| optimizer_config = OptimizerConfig( |
| optimizer='adam', bf16=True, use_distributed_optimizer=use_distributed_optimizer |
| ) |
| elif precision == 'fp32': |
| optimizer_config = OptimizerConfig( |
| optimizer='adam', |
| bf16=False, |
| fp16=False, |
| use_distributed_optimizer=use_distributed_optimizer, |
| ) |
| optim = get_megatron_optimizer(optimizer_config, [model]) |
|
|
| model_sharded_state_dict = model.sharded_state_dict() |
| sharded_state_dict = optim.sharded_state_dict(model_sharded_state_dict) |
|
|
| if 'optimizer' in sharded_state_dict and 'state' in sharded_state_dict['optimizer']: |
| assert ( |
| 'common_step' not in sharded_state_dict['optimizer']['state'] |
| or sharded_state_dict['optimizer']['state']['common_step'] is not None |
| ), "Found 'optimizer.state.common_step=None' in sharded state dict." |
|
|
|
|
| def test_optimizer_reload_model_params(): |
| world = int(os.getenv('WORLD_SIZE', '1')) |
| rank = int(os.getenv('RANK', '0')) |
| _init_distributed(world, rank) |
| Utils.initialize_model_parallel() |
|
|
| model = Net().bfloat16().cuda() |
| |
| for param in model.parameters(): |
| param.data.fill_(1.0) |
| ddp_config = DistributedDataParallelConfig(use_distributed_optimizer=True) |
| model = DistributedDataParallel( |
| TransformerConfig(num_attention_heads=1, num_layers=1), ddp_config, model |
| ) |
| optimizer_config = OptimizerConfig(optimizer='adam', bf16=True, use_distributed_optimizer=True) |
| optim = get_megatron_optimizer(optimizer_config, [model]) |
|
|
| |
| for param in model.parameters(): |
| param.data.fill_(2.0) |
|
|
| |
| |
| for group in optim.param_groups: |
| for main_param in group['params']: |
| assert main_param.dtype == torch.float32 |
| torch.testing.assert_close( |
| main_param, torch.empty_like(main_param).fill_(1.0), atol=0, rtol=0 |
| ) |
|
|
| |
| optim.reload_model_params() |
| for group in optim.param_groups: |
| for main_param in group['params']: |
| assert main_param.dtype == torch.float32 |
| torch.testing.assert_close( |
| main_param, torch.empty_like(main_param).fill_(2.0), atol=0, rtol=0 |
| ) |
|
|
| |
| state_dict = model.state_dict() |
| new_state_dict = {} |
| for name, param in state_dict.items(): |
| new_state_dict[name] = torch.empty_like(param).fill_(3.0) |
|
|
| |
| |
| optim.reload_model_params(new_state_dict) |
| for param in model.parameters(): |
| torch.testing.assert_close(param, torch.empty_like(param).fill_(2.0), atol=0, rtol=0) |
| for group in optim.param_groups: |
| for main_param in group['params']: |
| assert main_param.dtype == torch.float32 |
| torch.testing.assert_close( |
| main_param, torch.empty_like(main_param).fill_(3.0), atol=0, rtol=0 |
| ) |
|
|
|
|
| @pytest.mark.skipif( |
| not is_torch_min_version("2.4.0"), |
| reason="torch.distributed.init_device_mesh requires torch >= 2.4.0", |
| ) |
| @pytest.mark.parametrize( |
| "world_size, tp_size, cp_size, dp_size", |
| [ |
| (1, 1, 1, 1), |
| (2, 1, 2, 1), |
| (2, 2, 1, 1), |
| (8, 8, 1, 1), |
| (8, 2, 4, 1), |
| (8, 4, 2, 1), |
| (8, 1, 1, 8), |
| (8, 2, 1, 4), |
| (8, 2, 2, 2), |
| ], |
| ) |
| def test_get_megatron_optimizer_with_custom_process_groups(world_size, tp_size, cp_size, dp_size): |
| """ |
| Test that get_megatron_optimizer works correctly with custom process groups |
| provided via pg_collection parameters. |
| """ |
| |
| actual_world_size = torch.cuda.device_count() |
| if actual_world_size != world_size: |
| pytest.skip(f"Test requires world_size={world_size}, but got {actual_world_size}") |
|
|
| |
| Utils.initialize_model_parallel( |
| tensor_model_parallel_size=tp_size, context_parallel_size=cp_size |
| ) |
|
|
| |
| device_mesh = torch.distributed.init_device_mesh( |
| "cuda", (1, dp_size, 1, cp_size, tp_size), mesh_dim_names=("pp", "dp", "ep", "cp", "tp") |
| ) |
|
|
| |
| dp_group = device_mesh.get_group(mesh_dim="dp") |
| cp_group = device_mesh.get_group(mesh_dim="cp") |
| tp_group = device_mesh.get_group(mesh_dim="tp") |
| pp_group = device_mesh.get_group(mesh_dim="pp") |
|
|
| |
| dp_cp_mesh = device_mesh["dp", "cp"] |
| dp_cp_group = dp_cp_mesh._flatten().get_group() |
|
|
| |
| mp_mesh = device_mesh["pp", "tp"] |
| mp_group = mp_mesh._flatten().get_group() |
|
|
| |
| pg_collection = ProcessGroupCollection() |
| pg_collection.dp = dp_group |
| pg_collection.dp_cp = dp_cp_group |
| pg_collection.expt_dp = None |
|
|
| pg_collection.tp = tp_group |
| pg_collection.cp = cp_group |
| pg_collection.pp = pp_group |
| pg_collection.mp = mp_group |
| pg_collection.tp_ep_pp = None |
|
|
| |
| model = torch.nn.Linear(100, 100, bias=False, device='cuda') |
| model.requires_grad_(True) |
| model.weight.data.fill_(1.0) |
| ddp_config = DistributedDataParallelConfig(use_distributed_optimizer=True) |
| model = DistributedDataParallel( |
| TransformerConfig(num_attention_heads=1, num_layers=1), ddp_config, model |
| ) |
| for param in model.parameters(): |
| assert param.requires_grad |
| model_chunks = [model] |
|
|
| |
| optimizer_config = OptimizerConfig( |
| optimizer='adam', |
| lr=0.001, |
| weight_decay=0.01, |
| adam_beta1=0.9, |
| adam_beta2=0.999, |
| adam_eps=1e-8, |
| ) |
|
|
| |
| optimizer = get_megatron_optimizer( |
| config=optimizer_config, |
| model_chunks=model_chunks, |
| use_gloo_process_groups=False, |
| pg_collection=pg_collection, |
| ) |
|
|
| |
| assert optimizer is not None, "Optimizer should not be None" |
| assert hasattr(optimizer, 'param_groups'), "Optimizer should have param_groups" |
| assert len(optimizer.param_groups) > 0, "Optimizer should have at least one parameter group" |
|
|
| |
| input_tensor = torch.randn(32, 100, device='cuda', requires_grad=True) |
| output = model(input_tensor) |
| loss = output.sum() |
| loss.backward() |
|
|
| |
| optimizer.zero_grad() |
| output = model(input_tensor) |
| loss = output.sum() |
| loss.backward() |
|
|
| |
| original_weight = model.module.weight.data.clone() |
| original_bias = model.module.bias.data.clone() if model.module.bias is not None else None |
|
|
| |
| optimizer.step() |
|
|
| |
| assert not torch.equal( |
| model.module.weight.data, original_weight |
| ), "Weight should be updated after optimizer step" |
| if model.module.bias is not None: |
| assert not torch.equal( |
| model.module.bias.data, original_bias |
| ), "Bias should be updated after optimizer step" |
|
|
| |
| if world_size == 1: |
| |
| default_optimizer = get_megatron_optimizer( |
| config=optimizer_config, model_chunks=model_chunks |
| ) |
|
|
| |
| assert len(optimizer.param_groups) == len( |
| default_optimizer.param_groups |
| ), "Custom and default optimizers should have same number of parameter groups" |
|
|
|
|
| def test_get_megatron_optimizer_custom_process_groups_validation(): |
| """ |
| Test validation logic for custom process groups in get_megatron_optimizer. |
| """ |
| Utils.initialize_model_parallel(tensor_model_parallel_size=1) |
|
|
| |
| model = torch.nn.Linear(100, 100, bias=False, device='cuda') |
| model.requires_grad_(True) |
| model.weight.data.fill_(1.0) |
| ddp_config = DistributedDataParallelConfig(use_distributed_optimizer=True) |
| model = DistributedDataParallel( |
| TransformerConfig(num_attention_heads=1, num_layers=1), ddp_config, model |
| ) |
| for param in model.parameters(): |
| assert param.requires_grad |
| model_chunks = [model] |
| optimizer_config = OptimizerConfig(optimizer='adam', lr=0.001) |
|
|
| |
| pg_collection_no_dp = ProcessGroupCollection() |
|
|
| with pytest.raises(ValueError, match="dp process group is required"): |
| get_megatron_optimizer( |
| config=optimizer_config, model_chunks=model_chunks, pg_collection=pg_collection_no_dp |
| ) |
|
|
| |
| pg_collection_no_expt_dp = ProcessGroupCollection() |
| pg_collection_no_expt_dp.dp = torch.distributed.new_group() |
| |
|
|
| with pytest.raises(ValueError, match="expt_dp process group is required"): |
| get_megatron_optimizer( |
| config=optimizer_config, |
| model_chunks=model_chunks, |
| pg_collection=pg_collection_no_expt_dp, |
| ) |
|
|
| |
| pg_collection_complete = ProcessGroupCollection() |
| pg_collection_complete.dp = torch.distributed.new_group() |
| pg_collection_complete.expt_dp = None |
| |
|
|
| with pytest.raises(ValueError, match="mp process group is required"): |
| get_megatron_optimizer( |
| config=optimizer_config, model_chunks=model_chunks, pg_collection=pg_collection_complete |
| ) |
|
|
| |
| pg_collection_complete.mp = None |
|
|
| with pytest.raises(ValueError, match="tp_ep_pp process group is required"): |
| get_megatron_optimizer( |
| config=optimizer_config, model_chunks=model_chunks, pg_collection=pg_collection_complete |
| ) |
|
|
| |
| pg_collection_complete.mp = None |
| pg_collection_complete.tp_ep_pp = None |
|
|
| with pytest.raises(ValueError, match="Gloo process groups are not supported"): |
| get_megatron_optimizer( |
| config=optimizer_config, |
| model_chunks=model_chunks, |
| use_gloo_process_groups=True, |
| pg_collection=pg_collection_complete, |
| ) |
|
|