TransformerEngine Integration (#1282)

* Implemented ColumnParallelLinear with Transformer-Engine * Implemented RowParallelLinear with Transformer-Engine * Implemented LayerNormMLP with Transformer-Engine * Implemented MultiheadAttention with Transformer-Engine * Cleaned up transformer.py * Cleaned up neox_args * Cleaned up neox_args * - Fixed TE_MHA and added rope support - Implemented delayed scaling * Fixed mixed files. * Implemented ColumnParallelLinear with Transformer-Engine * Implemented RowParallelLinear with Transformer-Engine * Implemented LayerNormMLP with Transformer-Engine * Implemented MultiheadAttention with Transformer-Engine * Cleaned up transformer.py * Cleaned up neox_args * Cleaned up neox_args * - Fixed TE_MHA and added rope support - Implemented delayed scaling * Fixed mixed files. * Changed get_linear name * Added rng tracker to lnmlp and placed rope in te_mha init instead of forward * Updated fp8 arguments to te_fp8 * Added EAI copyright * precommit * add sample TE config * add te to readme * remove pip install prefix from reqs file * Force TE pytorch in requirements file --------- Co-authored-by: Quentin Anthony <[email protected]>
EleutherAI · Dec 19, 2024 · 8900d05 · 8900d05
1 parent 29080c3
commit 8900d05
Show file tree

Hide file tree

Showing 8 changed files with 968 additions and 171 deletions.
diff --git a/README.md b/README.md
@@ -18,6 +18,8 @@ GPT-NeoX leverages many of the same features and technologies as the popular Meg
 * Easy connections with the open source ecosystem, including Hugging Face's [tokenizers](https://github.com/huggingface/tokenizers) and [transformers](https://github.com/huggingface/transformers/) libraries, monitor experiments via [WandB](https://wandb.ai/site)/[Comet](https://www.comet.com/site/)/TensorBoard, and evaluation via our [Language Model Evaluation Harness](https://github.com/EleutherAI/lm-evaluation-harness).
 
 ## News
+**[10/9/2024]** We now support [Transformer Engine](https://github.com/NVIDIA/TransformerEngine) integration
+
 **[9/9/2024]** We now support preference learning via [DPO](https://arxiv.org/abs/2305.18290), [KTO](https://arxiv.org/abs/2402.01306), and reward modeling
 
 **[9/9/2024]** We now support integration with [Comet ML](https://www.comet.com/site/), a machine learning monitoring platform
@@ -60,6 +62,7 @@ Prior to 3/9/2023, GPT-NeoX relied on [DeeperSpeed](https://github.com/EleutherA
   * [Environment and Dependencies](#environment-and-dependencies)
     + [Host Setup](#host-setup)
     + [Flash Attention](#flash-attention)
+    + [Transformer Engine](#transformer-engine)
     + [Multi-Node Launching](#multi-node-launching)
     + [Containerized Setup](#containerized-setup)
   * [Usage](#usage)
@@ -130,7 +133,20 @@ This will automatically adapts building process over different GPU vendors (AMD,
 
 ### Flash Attention
 
-To use [Flash-Attention](https://github.com/HazyResearch/flash-attention), install the additional dependencies in  `./requirements/requirements-flashattention.txt` and set the attention type in your configuration accordingly (see [configs](./configs/)). This can provide significant speed-ups over regular attention on certain GPU architectures, including Ampere GPUs (such as A100s); see the repository for more details.
+To use [Flash-Attention](https://github.com/HazyResearch/flash-attention), install the additional dependencies in  `./requirements/requirements-flashattention.txt` or use a PyTorch NGC container with it pre-installed (note that functionality is not guaranteed using versions different from our requirements file). Then set the attention type in your configuration accordingly (see [configs](./configs/)). This can provide significant speed-ups over regular attention on certain GPU architectures, including Ampere GPUs (such as A100s); see the repository for more details.
+
+### Transformer Engine
+
+To use [Transformer Engine (TE)](https://github.com/NVIDIA/TransformerEngine), install the additional dependencies in  `./requirements/requirements-transformer-engine.txt` or use a PyTorch NGC container with it pre-installed (note that functionality is not guaranteed using versions different from our requirements file). See [this config](https://github.com/EleutherAI/gpt-neox/configs/1-3B-transformer-engine.yml) for an example of using TE on a 1.3B model. This can provide significant speed-ups over regular attention on certain GPU architectures, including Ampere and Hopper GPUs; see the repository for more details.
+
+
+TE provides very efficient kernels for both A100 and H100 GPUs. We've run some sample ablations on A100:
+
+
+
+and H100:
+
+
 
 
 ### Multi-Node Launching

diff --git a/configs/1-3B-transformer-engine.yml b/configs/1-3B-transformer-engine.yml
@@ -0,0 +1,105 @@
+# GPT-2 pretraining setup
+{
+   # parallelism settings ( you will want to change these based on your cluster setup, ideally scheduling pipeline stages
+   # across the node boundaries )
+   "pipe_parallel_size": 1,
+   "model_parallel_size": 1,
+
+   # model settings
+   "num_layers": 24,
+   "hidden_size": 2048,
+   "num_attention_heads": 16,
+   "seq_length": 2048,
+   "max_position_embeddings": 2048,
+   "norm": "layernorm",
+   "pos_emb": "rotary",
+   "no_weight_tying": true,
+   "gpt_j_residual": false,
+   "output_layer_parallelism": "column",
+
+   # Transformer Engine settings
+   "te_columnparallel": false,
+   "te_rowparallel": false,
+   "te_layernorm_mlp": true,
+   "te_mha": true,
+   "te_fp8_format": "hybrid",
+   "te_fp8_wgrad": true,
+   "te_fp8_amax_history_len": 1,
+   "te_fp8_amax_compute_algo": "most_recent",
+   "te_fp8_margin": 0,
+   "te_fp8_mha": false,
+
+   # these should provide some speedup but takes a while to build, set to true if desired
+   "scaled_upper_triang_masked_softmax_fusion": false,
+   "bias_gelu_fusion": false,
+   "rope_fusion": false,
+   "layernorm_fusion": false,
+
+   # init methods
+   "init_method": "small_init",
+   "output_layer_init_method": "wang_init",
+
+   # optimizer settings
+   "optimizer": {
+     "type": "Adam",
+     "params": {
+       "lr": 0.0002,
+       "betas": [0.9, 0.95],
+       "eps":  1.0e-8,
+     }
+   },
+   "min_lr": 0.00002,
+
+   # for all zero_optimization options, see https://www.deepspeed.ai/docs/config-json/#zero-optimizations-for-fp16-training
+   "zero_optimization": {
+    "stage": 1,
+    "allgather_partitions": True,
+    "allgather_bucket_size": 500000000,
+    "overlap_comm": True,
+    "reduce_scatter": True,
+    "reduce_bucket_size": 500000000,
+    "contiguous_gradients": True,
+  },
+
+   # batch / data settings
+   "train_micro_batch_size_per_gpu": 4,
+   "data_impl": "mmap",
+
+   # activation checkpointing
+   "checkpoint_activations": true,
+   "checkpoint_num_layers": 1,
+   "partition_activations": true,
+   "synchronize_each_layer": true,
+
+   # regularization
+   "gradient_clipping": 1.0,
+   "weight_decay": 0.1,
+   "hidden_dropout": 0,
+   "attention_dropout": 0,
+
+   # precision settings
+   "fp16": {
+     "fp16": true,
+     "enabled": true,
+     "loss_scale": 0,
+     "loss_scale_window": 1000,
+     "hysteresis": 2,
+     "min_loss_scale": 1
+   },
+
+   # misc. training settings
+   "train_iters": 320000,
+   "lr_decay_iters": 320000,
+   "distributed_backend": "nccl",
+   "lr_decay_style": "cosine",
+   "warmup": 0.01,
+   "checkpoint_factor": 10000,
+   "eval_interval": 1000,
+   "eval_iters": 10,
+
+   # logging
+   "log_interval": 100,
+   "steps_per_print": 10,
+   "keep_last_n_checkpoints": 4,
+   "wall_clock_breakdown": true,
+}
diff --git a/megatron/model/positional_embeddings.py b/megatron/model/positional_embeddings.py
@@ -67,6 +67,8 @@ def _prepare_cache(self, seq_len, precision, base):
         freqs = torch.einsum("i,j->ij", t, inv_freq)
         emb = torch.cat((freqs, freqs), dim=-1)
 
+        self.emb = emb.reshape(emb.size(0), 1, 1, emb.size(1))
+
         cos_cached = emb.cos()[:, None, None, :]
         sin_cached = emb.sin()[:, None, None, :]
 
@@ -76,6 +78,9 @@ def _prepare_cache(self, seq_len, precision, base):
             inv_freq.to(precision),
         )
 
+    def get_emb(self):
+        return self.emb.to(self.precision).cuda()
+
     def forward(self, x, seq_dim=0, seq_len=None):
         if seq_len is None:
             seq_len = x.shape[seq_dim]