Inference: general tensor-parallel examples (#144)

* add T5 example using tensor-parallelism * refine t5 test * add more tests to try the Tensor-Parallel inference * remove pdb * add tests for GPT-J and wav2vec2 model architectures Co-authored-by: Ammar Ahmad Awan <[email protected]>
microsoft · Feb 22, 2022 · 9c48e36 · 9c48e36
1 parent 068e656
commit 9c48e36
Show file tree

Hide file tree

Showing 6 changed files with 174 additions and 2 deletions.
diff --git a/inference/huggingface/gpt-neo.py b/inference/huggingface/gpt-neo.py
@@ -15,7 +15,6 @@
     "***************** Creating model in RANK ({0}) with WORLD_SIZE = {1} *****************"
     .format(local_rank,
             world_size))
-
 generator = pipeline('text-generation',
                      model='EleutherAI/gpt-neo-2.7B',
                      device=local_rank)
@@ -25,4 +24,4 @@
                                            replace_method='auto',
                                            replace_with_kernel_inject=True)
 string = generator("DeepSpeed is", do_sample=True, min_length=50)
-print(string)
+print(string)
diff --git a/inference/huggingface/test-electra.py b/inference/huggingface/test-electra.py
@@ -0,0 +1,28 @@
+from transformers import pipeline
+import transformers
+import deepspeed
+import torch
+import os
+from transformers.models.electra.modeling_electra import ElectraLayer
+
+local_rank = int(os.getenv('LOCAL_RANK', '0'))
+world_size = int(os.getenv('WORLD_SIZE', '4'))
+
+pipe = pipeline('fill-mask', model="google/electra-base-generator",
+    tokenizer="google/electra-base-generator")
+
+# The inpjection_policy shows two things:
+#   1. which layer module we need to add Tensor-Parallelism
+#   2. the name of one or several linear layers: a) attention_output (both encoder and decoder), 
+#       and b) transformer output
+pipe.model = deepspeed.init_inference(
+    pipe.model,
+    mp_size=world_size,
+    dtype=torch.float,
+    injection_policy={ElectraLayer: ('output.dense')}
+)
+pipe.device = torch.device(f'cuda:{local_rank}')
+output = pipe(f"HuggingFace is creating a {pipe.tokenizer.mask_token} that the community uses to solve NLP tasks.")
+
+if not torch.distributed.is_initialized() or torch.distributed.get_rank() == 0:
+    print(output)
diff --git a/inference/huggingface/test-gptj.py b/inference/huggingface/test-gptj.py
@@ -0,0 +1,36 @@
+import os
+import torch
+import deepspeed
+import transformers
+
+from deepspeed import module_inject
+from transformers import pipeline
+from transformers.models.gptj.modeling_gptj import GPTJBlock
+
+# Get local gpu rank from torch.distributed/deepspeed launcher
+local_rank = int(os.getenv('LOCAL_RANK', '0'))
+world_size = int(os.getenv('WORLD_SIZE', '1'))
+
+print(
+    "***************** Creating model in RANK ({0}) with WORLD_SIZE = {1} *****************"
+    .format(local_rank,
+            world_size))
+from transformers import AutoTokenizer, AutoModelForCausalLM
+
+tokenizer = AutoTokenizer.from_pretrained("EleutherAI/gpt-j-6B")
+model = AutoModelForCausalLM.from_pretrained("EleutherAI/gpt-j-6B")
+
+inp_tokens = tokenizer("DeepSpeed is", return_tensors="pt",)
+model = deepspeed.init_inference(model,
+                                 mp_size=world_size,
+                                 dtype=torch.float,
+                                 injection_policy={GPTJBlock: ('attn.out_proj','mlp.fc_out')},
+                                 replace_with_kernel_inject=False)
+
+for token in inp_tokens:
+    if torch.is_tensor(inp_tokens[token]):
+        inp_tokens[token] = inp_tokens[token].to(f'cuda:{local_rank}')
+
+model.cuda().to(f'cuda:{local_rank}')
+string = tokenizer.batch_decode(model.generate(**inp_tokens,min_length=50,))[0]
+print(string)
diff --git a/inference/huggingface/test-roberta.py b/inference/huggingface/test-roberta.py
@@ -0,0 +1,29 @@
+from transformers import pipeline
+import transformers
+import deepspeed
+import torch
+import os
+from transformers.models.roberta.modeling_roberta import RobertaLayer
+
+local_rank = int(os.getenv('LOCAL_RANK', '0'))
+world_size = int(os.getenv('WORLD_SIZE', '4'))
+
+pipe = pipeline('fill-mask', model="roberta-large", device=local_rank)
+
+# The inpjection_policy shows two things:
+#   1. which layer module we need to add Tensor-Parallelism
+#   2. the name of several linear layers: a) attention_output (both encoder and decoder), 
+#       and b) transformer output
+
+pipe.model = deepspeed.init_inference(
+    pipe.model,
+    mp_size=world_size,
+    dtype=torch.float,
+    injection_policy={RobertaLayer: ('output.dense')}
+)
+
+pipe.device = torch.device(f'cuda:{local_rank}')
+output = pipe("Hello I'm a <mask> model.")
+
+if not torch.distributed.is_initialized() or torch.distributed.get_rank() == 0:
+    print(output)
diff --git a/inference/huggingface/test-t5.py b/inference/huggingface/test-t5.py
@@ -0,0 +1,29 @@
+from transformers import pipeline
+import transformers
+import deepspeed
+import torch
+import os
+from transformers.models.t5.modeling_t5 import T5Block
+
+local_rank = int(os.getenv('LOCAL_RANK', '0'))
+world_size = int(os.getenv('WORLD_SIZE', '4'))
+
+pipe = pipeline("text2text-generation", model="google/t5-v1_1-small", device=local_rank)
+
+# The inpjection_policy shows two things:
+#   1. which layer module we need to add Tensor-Parallelism
+#   2. the name of several linear layers: a) attention_output (both encoder and decoder), 
+#       and b) transformer output
+
+pipe.model = deepspeed.init_inference(
+    pipe.model,
+    mp_size=world_size,
+    dtype=torch.float,
+    injection_policy={T5Block: ('SelfAttention.o', 'EncDecAttention.o', 'DenseReluDense.wo')}
+)
+
+pipe.device = torch.device(f'cuda:{local_rank}')
+output = pipe("Is this review positive or negative? Review: this is the best cast iron skillet you will ever buy")
+
+if not torch.distributed.is_initialized() or torch.distributed.get_rank() == 0:
+    print(output)
diff --git a/inference/huggingface/test-wav2vec2.py b/inference/huggingface/test-wav2vec2.py
@@ -0,0 +1,51 @@
+from datasets import load_dataset
+from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
+import soundfile as sf
+import torch
+from jiwer import wer
+import os
+import torch
+import deepspeed
+from deepspeed import module_inject
+from transformers.models.wav2vec2.modeling_wav2vec2 import Wav2Vec2EncoderLayer
+
+librispeech_eval = load_dataset("librispeech_asr", "clean", split="test")
+
+# Get local gpu rank from torch.distributed/deepspeed launcher
+local_rank = int(os.getenv('LOCAL_RANK', '0'))
+world_size = int(os.getenv('WORLD_SIZE', '1'))
+
+print(
+    "***************** Creating model in RANK ({0}) with WORLD_SIZE = {1} *****************"
+    .format(local_rank,
+            world_size))
+
+model = Wav2Vec2ForCTC.from_pretrained("facebook/wav2vec2-base-960h")
+processor = Wav2Vec2Processor.from_pretrained("facebook/wav2vec2-base-960h")
+
+model = deepspeed.init_inference(model,
+                                 mp_size=world_size,
+                                 dtype=torch.float,
+                                 injection_policy={Wav2Vec2EncoderLayer: ('attention.out_proj','feed_forward.output_dense')},
+                                 replace_with_kernel_inject=False)
+model.to(f'cuda:{local_rank}')                                 
+def map_to_array(batch):
+    speech, _ = sf.read(batch["file"])
+    batch["speech"] = speech
+    return batch
+
+librispeech_eval = librispeech_eval.map(map_to_array)
+
+def map_to_pred(batch):
+    input_values = processor(batch["speech"], return_tensors="pt", padding="longest").input_values
+    with torch.no_grad():
+        logits = model(input_values.to(f'cuda:{local_rank}')).logits
+
+    predicted_ids = torch.argmax(logits, dim=-1)
+    transcription = processor.batch_decode(predicted_ids)
+    batch["transcription"] = transcription
+    return batch
+
+result = librispeech_eval.map(map_to_pred, batched=True, batch_size=1, remove_columns=["speech"])
+
+print("WER:", wer(result["text"], result["transcription"]))