[inference] add reference and fix some bugs (#4937)

* add reference and fix some bugs

* update gptq init

---------
Co-authored-by: Xu Kai <xukai16@foxamil.com>

colossalai/inference/quant/smoothquant/models/base_model.py

@@ -132,6 +132,7 @@ class BaseSmoothForCausalLM(nn.Module, PushToHubMixin):
             mean_scale = np.mean([v["input"] for v in act_dict.values()])
             pbar.set_description(f"Mean input scale: {mean_scale:.2f}")
 
+    # Adatped from https://github.com/mit-han-lab/smoothquant/blob/main/smoothquant/calibration.py
     def get_act_scales(self, model, tokenizer, dataset, num_samples=512, seq_len=512):
         model.eval()
         device = next(model.parameters()).device
@@ -163,6 +164,7 @@ class BaseSmoothForCausalLM(nn.Module, PushToHubMixin):
 
         return act_scales
 
+    # Adapted from https://github.com/mit-han-lab/smoothquant/blob/main/smoothquant/smooth.py
     @torch.no_grad()
     def smooth_ln_fcs(self, ln, fcs, act_scales, alpha=0.5):
         if not isinstance(fcs, list):
@@ -189,6 +191,7 @@ class BaseSmoothForCausalLM(nn.Module, PushToHubMixin):
     def create_quantized_model(model):
         raise NotImplementedError("Not implement create_quantized_model method")
 
+    # Adapted from AutoGPTQ: https://github.com/PanQiWei/AutoGPTQ/blob/main/auto_gptq/modeling/_base.py
     def save_quantized(
         self,
         save_dir: str,
@@ -249,6 +252,7 @@ class BaseSmoothForCausalLM(nn.Module, PushToHubMixin):
 
         self.model.config.save_pretrained(save_dir)
 
+    # Adapted from AutoGPTQ: https://github.com/PanQiWei/AutoGPTQ/blob/main/auto_gptq/modeling/_base.py
     def save_pretrained(
         self,
         save_dir: str,
@@ -260,6 +264,7 @@ class BaseSmoothForCausalLM(nn.Module, PushToHubMixin):
         warnings.warn("you are using save_pretrained, which will re-direct to save_quantized.")
         self.save_quantized(save_dir, use_safetensors, safetensors_metadata)
 
+    # Adapted from AutoGPTQ: https://github.com/PanQiWei/AutoGPTQ/blob/main/auto_gptq/modeling/_base.py
     @classmethod
     def from_pretrained(
         cls,
@@ -354,6 +359,7 @@ class BaseSmoothForCausalLM(nn.Module, PushToHubMixin):
 
         return cls(model, False)
 
+    # Adapted from AutoGPTQ: https://github.com/PanQiWei/AutoGPTQ/blob/main/auto_gptq/modeling/_base.py
     @classmethod
     def from_quantized(
         cls,

colossalai/inference/quant/smoothquant/models/linear.py

@@ -62,6 +62,7 @@ class W8A8BFP32O32LinearSiLU(torch.nn.Module):
         return int8_module
 
 
+# modified from torch-int: https://github.com/Guangxuan-Xiao/torch-int/blob/main/torch_int/nn/linear.py
 class W8A8B8O8Linear(torch.nn.Module):
     # For qkv_proj
     def __init__(self, in_features, out_features, alpha=1.0, beta=1.0):
@@ -117,6 +118,7 @@ class W8A8B8O8Linear(torch.nn.Module):
         return int8_module
 
 
+# modified from torch-int: https://github.com/Guangxuan-Xiao/torch-int/blob/main/torch_int/nn/linear.py
 class W8A8BFP32OFP32Linear(torch.nn.Module):
     # For fc2 and out_proj
     def __init__(self, in_features, out_features, alpha=1.0, beta=1.0):

colossalai/inference/quant/smoothquant/models/llama.py

@@ -419,6 +419,7 @@ class LlamaApplyRotary(nn.Module):
         return x_embed
 
 
+# Adapted from https://github.com/huggingface/transformers/blob/main/src/transformers/models/llama/modeling_llama.py
 def llama_decoder_layer_forward(
     self,
     hidden_states: torch.Tensor,
@@ -559,6 +560,7 @@ def init_to_get_rotary(config, base=10000, use_elem=False):
     return _cos_cached, _sin_cached
 
 
+# Adapted from https://github.com/huggingface/transformers/blob/main/src/transformers/models/llama/modeling_llama.py
 @add_start_docstrings_to_model_forward(LLAMA_INPUTS_DOCSTRING)
 def llama_model_forward(
     self,
@@ -729,6 +731,7 @@ class SmoothLlamaForCausalLM(BaseSmoothForCausalLM):
     def __init__(self, model: PreTrainedModel, quantized: bool = False):
         super().__init__(model, quantized)
 
+    # Adatped from https://github.com/mit-han-lab/smoothquant/blob/main/smoothquant/calibration.py
     def get_act_dict(
         self,
         tokenizer,

colossalai/inference/tensor_parallel/engine.py

@@ -21,6 +21,8 @@ _supported_models = [
     "BloomForCausalLM",
     "ChatGLMModel",
     "ChatGLMForConditionalGeneration",
+    "LlamaGPTQForCausalLM",
+    "BloomGPTQForCausalLM",
 ]
 
 
@@ -213,11 +215,14 @@ class TPInferEngine:
         ), "Discrepancy between the tp size of TPInferEngine and the tp size of shard config"
         model_name = model.__class__.__name__
         assert model_name in self.supported_models, f"Unsupported model cls {model_name} for TP inference."
+
+        model = model.model if self.shard_config.inference_gptq else model
+
         policy = get_autopolicy(model, inference_only=True)
         self.model, _ = shardformer.optimize(model, policy)
 
         if self.shard_config.inference_gptq:
-            self._post_init_gptq_buffer(model)
+            self._post_init_gptq_buffer(self.model)
 
         self.model = self.model.cuda()
 

colossalai/kernel/triton/gptq_triton.py

@@ -267,6 +267,7 @@ def cai_gptq_matmul_248_kernel(
     tl.store(c_ptrs, accumulator, mask=c_mask)
 
 
+# Adapted from AutoGPTQ auto_gptq: https://github.com/PanQiWei/AutoGPTQ
 @autotune(
     configs=[
         triton.Config(

colossalai/kernel/triton/smooth_attention.py

@@ -13,10 +13,10 @@ except ImportError:
 
 if HAS_TRITON:
     """
-    this function is modified from
-    https://github.com/ModelTC/lightllm/blob/f093edc20683ac3ea1bca3fb5d8320a0dd36cf7b/lightllm/models/llama/triton_kernel/context_flashattention_nopad.py#L10
+    this functions are modified from https://github.com/ModelTC/lightllm
     """
 
+    # Adapted from  https://github.com/ModelTC/lightllm/blob/main/lightllm/models/llama/triton_kernel/context_flashattention_nopad.py
     @triton.jit
     def _context_flash_attention_kernel(
         Q,
@@ -145,20 +145,16 @@ if HAS_TRITON:
         tl.store(out_ptrs, acc, mask=offs_m[:, None] < cur_batch_seq_len)
         return
 
-
-
     @torch.no_grad()
     def smooth_llama_context_attn_fwd(
         q, k, v, o, q_input_scale, k_input_scale, v_input_scale, pv_output_scale, b_start_loc, b_seq_len, max_input_len
     ):
-
         BLOCK = 128
         # shape constraints
         Lq, Lk, Lv = q.shape[-1], k.shape[-1], v.shape[-1]
         assert Lq == Lk, "context process only supports equal query, key, value length"
         assert Lk == Lv, "context process only supports equal query, key, value length"
         assert Lk in {16, 32, 64, 128}
-        BLOCK_N = 128
         sm_scale = 1.0 / math.sqrt(Lk)
         batch, head = b_seq_len.shape[0], q.shape[1]
         grid = (batch, head, triton.cdiv(max_input_len, BLOCK))
@@ -203,6 +199,7 @@ if HAS_TRITON:
         )
         return
 
+    # Adapted from https://github.com/ModelTC/lightllm/blob/main/lightllm/models/llama/triton_kernel/token_attention_nopad_att1.py
     @triton.jit
     def _token_attn_1_kernel(
         Q,
@@ -264,6 +261,7 @@ if HAS_TRITON:
             tl.store(attn_out + off_o, att_value, mask=offs_n_new < current_batch_end_index)
         return
 
+    # Adapted from https://github.com/ModelTC/lightllm/blob/main/lightllm/models/llama/triton_kernel/token_attention_nopad_att1.py
     @triton.jit
     def _token_attn_1_alibi_kernel(
         Q,
@@ -413,6 +411,7 @@ if HAS_TRITON:
             )
         return
 
+    # Adapted from https://github.com/ModelTC/lightllm/blob/main/lightllm/models/llama/triton_kernel/token_attention_softmax_and_reducev.py
     @triton.jit
     def _token_attn_softmax_fwd(
         softmax_logics,
@@ -479,6 +478,7 @@ if HAS_TRITON:
         )
         return
 
+    # Adapted from https://github.com/ModelTC/lightllm/blob/main/lightllm/models/llama/triton_kernel/token_attention_nopad_att1.py
     @triton.jit
     def _token_attn_2_kernel(
         Prob,

examples/inference/gptq_llama.py

@@ -8,7 +8,6 @@ from transformers import LlamaTokenizer
 
 import colossalai
 from colossalai.inference.tensor_parallel.engine import TPInferEngine
-from colossalai.inference.tensor_parallel.modeling._utils import init_to_get_rotary
 from colossalai.logging import disable_existing_loggers
 from colossalai.shardformer import ShardConfig
 from colossalai.testing import clear_cache_before_run, rerun_if_address_is_in_use, spawn
@@ -50,8 +49,6 @@ def run_llama_test(args):
         quantized_model_dir, device=torch.cuda.current_device(), inject_fused_attention=False
     )
 
-    init_to_get_rotary(model.model.model, base=10000)
-
     model_config = model.config
     shard_config = ShardConfig(
         enable_tensor_parallelism=True if args.tp_size > 1 else False, inference_only=True, inference_gptq=True