Add FusedLinerCrossEntropy support for Phi3

README.md

@@ -193,7 +193,7 @@ loss.backward()
 | Gemma1      | `liger_kernel.transformers.apply_liger_kernel_to_gemma`    | RoPE, RMSNorm, GeGLU, CrossEntropyLoss, FusedLinearCrossEntropy         |
 | Gemma2      | `liger_kernel.transformers.apply_liger_kernel_to_gemma2`   | RoPE, RMSNorm, GeGLU, CrossEntropyLoss         |
 | Qwen2       | `liger_kernel.transformers.apply_liger_kernel_to_qwen2`    | RoPE, RMSNorm, SwiGLU, CrossEntropyLoss, FusedLinearCrossEntropy        |
-| Phi3        | `liger_kernel.transformers.apply_liger_kernel_to_phi3`     | RoPE, RMSNorm, SwiGLU, CrossEntropyLoss         |
+| Phi3        | `liger_kernel.transformers.apply_liger_kernel_to_phi3`     | RoPE, RMSNorm, SwiGLU, CrossEntropyLoss, FusedLinearCrossEntropy         |
 
 
 

src/liger_kernel/transformers/model/phi3.py

@@ -0,0 +1,136 @@
+from typing import List, Optional, Tuple, Union
+
+import torch
+from torch.nn import CrossEntropyLoss
+from transformers.modeling_outputs import CausalLMOutputWithPast
+from transformers.models.phi3.modeling_phi3 import (
+    _CONFIG_FOR_DOC,
+    PHI3_INPUTS_DOCSTRING,
+)
+from transformers.utils import (
+    add_start_docstrings_to_model_forward,
+    replace_return_docstrings,
+)
+
+from liger_kernel.transformers.fused_linear_cross_entropy import (
+    LigerFusedLinearCrossEntropyLoss,
+)
+
+
+@add_start_docstrings_to_model_forward(PHI3_INPUTS_DOCSTRING)
+@replace_return_docstrings(
+    output_type=CausalLMOutputWithPast, config_class=_CONFIG_FOR_DOC
+)
+def lce_forward(
+    self,
+    input_ids: torch.LongTensor = None,
+    attention_mask: Optional[torch.Tensor] = None,
+    position_ids: Optional[torch.LongTensor] = None,
+    past_key_values: Optional[List[torch.FloatTensor]] = None,
+    inputs_embeds: Optional[torch.FloatTensor] = None,
+    labels: Optional[torch.LongTensor] = None,
+    use_cache: Optional[bool] = None,
+    output_attentions: Optional[bool] = None,
+    output_hidden_states: Optional[bool] = None,
+    return_dict: Optional[bool] = None,
+    cache_position: Optional[torch.LongTensor] = None,
+) -> Union[Tuple, CausalLMOutputWithPast]:
+    r"""
+    Copy paste phi3 forward from transfomers v4.44.2 but replace torch cross entropy with liger fused linear cross entropy
+
+
+    Args:
+        labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
+            Labels for computing the masked language modeling loss. Indices should either be in `[0, ...,
+            config.vocab_size]` or -100 (see `input_ids` docstring). Tokens with indices set to `-100` are ignored
+            (masked), the loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]`.
+
+    Returns:
+
+    Example:
+
+    ```python
+    >>> from transformers import AutoTokenizer, Phi3ForCausalLM
+
+    >>> model = Phi3ForCausalLM.from_pretrained("microsoft/phi-3-mini-4k-instruct")
+    >>> tokenizer = AutoTokenizer.from_pretrained("microsoft/phi-3-mini-4k-instruct")
+
+    >>> prompt = "This is an example script ."
+    >>> inputs = tokenizer(prompt, return_tensors="pt")
+
+    >>> # Generate
+    >>> generate_ids = model.generate(inputs.input_ids, max_length=30)
+    >>> tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
+    'This is an example script .\n Certainly! Below is a sample script that demonstrates a simple task, such as calculating the sum'
+    ```"""
+
+    output_attentions = (
+        output_attentions
+        if output_attentions is not None
+        else self.config.output_attentions
+    )
+    output_hidden_states = (
+        output_hidden_states
+        if output_hidden_states is not None
+        else self.config.output_hidden_states
+    )
+    return_dict = (
+        return_dict if return_dict is not None else self.config.use_return_dict
+    )
+
+    # decoder outputs consists of (dec_features, layer_state, dec_hidden, dec_attn)
+    outputs = self.model(
+        input_ids=input_ids,
+        attention_mask=attention_mask,
+        position_ids=position_ids,
+        past_key_values=past_key_values,
+        inputs_embeds=inputs_embeds,
+        use_cache=use_cache,
+        output_attentions=output_attentions,
+        output_hidden_states=output_hidden_states,
+        return_dict=return_dict,
+    )
+
+    hidden_states = outputs[0]
+
+    loss = None
+    logits = None
+
+    if self.training and labels is not None:
+        shift_hidden_states = hidden_states[..., :-1, :].contiguous()
+        shift_labels = labels[..., 1:].contiguous()
+
+        # flatten tokens
+        shift_hidden_states = shift_hidden_states.view(-1, self.config.hidden_size)
+        shift_labels = shift_labels.view(-1)
+
+        lce = LigerFusedLinearCrossEntropyLoss()
+        loss = lce(self.lm_head.weight, shift_hidden_states, shift_labels)
+    else:
+        logits = self.lm_head(hidden_states)
+        logits = logits.float()
+
+        loss = None
+        if labels is not None:
+            # Shift so that tokens < n predict n
+            shift_logits = logits[..., :-1, :].contiguous()
+            shift_labels = labels[..., 1:].contiguous()
+            # Flatten the tokens
+            loss_fct = CrossEntropyLoss()
+            shift_logits = shift_logits.view(-1, self.config.vocab_size)
+            shift_labels = shift_labels.view(-1)
+            # Enable model parallelism
+            shift_labels = shift_labels.to(shift_logits.device)
+            loss = loss_fct(shift_logits, shift_labels)
+
+    if not return_dict:
+        output = (logits,) + outputs[1:]
+        return (loss,) + output if loss is not None else output
+
+    return CausalLMOutputWithPast(
+        loss=loss,
+        logits=logits,
+        past_key_values=outputs.past_key_values,
+        hidden_states=outputs.hidden_states,
+        attentions=outputs.attentions,
+    )

src/liger_kernel/transformers/monkey_patch.py

@@ -5,8 +5,9 @@
 from liger_kernel.transformers.cross_entropy import LigerCrossEntropyLoss
 from liger_kernel.transformers.geglu import LigerGEGLUMLP
 from liger_kernel.transformers.model.gemma import lce_forward as gemma_lce_forward
-from liger_kernel.transformers.model.llama import lce_forward
+from liger_kernel.transformers.model.llama import lce_forward as llama_lce_forward
 from liger_kernel.transformers.model.mistral import lce_forward as mistral_lce_forward
+from liger_kernel.transformers.model.phi3 import lce_forward as phi3_lce_forward
 from liger_kernel.transformers.model.qwen2 import lce_forward as qwen2_lce_forward
 from liger_kernel.transformers.rms_norm import LigerRMSNorm
 from liger_kernel.transformers.rope import liger_rotary_pos_emb
@@ -33,7 +34,7 @@ def apply_liger_kernel_to_llama(
         rope (bool): Whether to apply Liger's rotary position embedding. Default is True.
         cross_entropy (bool): Whether to apply Liger's cross entropy loss. Default is False.
         fused_linear_cross_entropy (bool):
-            Whether to apply Liger's fused lienar cross entropy loss. Default is True.
+            Whether to apply Liger's fused linear cross entropy loss. Default is True.
             `cross_entropy` and `fused_linear_cross_entropy` cannot both be True.
             If `fused_linear_cross_entropy` is True, the logits will not be materialized but more memory efficient.
         rms_norm (bool): Whether to apply Liger's RMSNorm. Default is True.
@@ -55,7 +56,7 @@ def apply_liger_kernel_to_llama(
     if cross_entropy:
         modeling_llama.CrossEntropyLoss = LigerCrossEntropyLoss
     if fused_linear_cross_entropy:
-        modeling_llama.LlamaForCausalLM.forward = lce_forward
+        modeling_llama.LlamaForCausalLM.forward = llama_lce_forward
 
 
 def apply_liger_kernel_to_mistral(
@@ -72,7 +73,7 @@ def apply_liger_kernel_to_mistral(
         rope (bool): Whether to apply Liger's rotary position embedding. Default is True.
         cross_entropy (bool): Whether to apply Liger's cross entropy loss. Default is True.
         fused_linear_cross_entropy (bool):
-            Whether to apply Liger's fused lienar cross entropy loss. Default is True.
+            Whether to apply Liger's fused linear cross entropy loss. Default is True.
             `cross_entropy` and `fused_linear_cross_entropy` cannot both be True.
             If `fused_linear_cross_entropy` is True, the logits will not be materialized but more memory efficient.
         rms_norm (bool): Whether to apply Liger's RMSNorm. Default is True.
@@ -206,7 +207,7 @@ def apply_liger_kernel_to_qwen2(
         rope (bool): Whether to apply Liger's rotary position embedding. Default is True.
         cross_entropy (bool): Whether to apply Liger's cross entropy loss. Default is False.
         fused_linear_cross_entropy (bool):
-            Whether to apply Liger's fused lienar cross entropy loss. Default is True.
+            Whether to apply Liger's fused linear cross entropy loss. Default is True.
             `cross_entropy` and `fused_linear_cross_entropy` cannot both be True.
             If `fused_linear_cross_entropy` is True, the logits will not be materialized but more memory efficient.
         rms_norm (bool): Whether to apply Liger's RMSNorm. Default is True.
@@ -232,7 +233,8 @@ def apply_liger_kernel_to_qwen2(
 
 def apply_liger_kernel_to_phi3(
     rope: bool = True,
-    cross_entropy: bool = True,
+    cross_entropy: bool = False,
+    fused_linear_cross_entropy: bool = True,
     rms_norm: bool = True,
     swiglu: bool = True,
 ) -> None:
@@ -242,9 +244,17 @@ def apply_liger_kernel_to_phi3(
     Args:
         rope (bool): Whether to apply Liger's rotary position embedding. Default is True.
         cross_entropy (bool): Whether to apply Liger's cross entropy loss. Default is False.
+        fused_linear_cross_entropy (bool):
+            Whether to apply Liger's fused linear cross entropy loss. Default is True.
+            `cross_entropy` and `fused_linear_cross_entropy` cannot both be True.
+            If `fused_linear_cross_entropy` is True, the logits will not be materialized but more memory efficient.
         rms_norm (bool): Whether to apply Liger's RMSNorm. Default is True.
         swiglu (bool): Whether to apply Liger's SwiGLU Phi3MLP. Default is True.
     """
+    assert not (
+        cross_entropy and fused_linear_cross_entropy
+    ), "cross_entropy and fused_linear_cross_entropy cannot both be True."
+
     from transformers.models.phi3 import modeling_phi3
 
     if rope:
@@ -255,11 +265,14 @@ def apply_liger_kernel_to_phi3(
         modeling_phi3.Phi3MLP = LigerPhi3SwiGLUMLP
     if cross_entropy:
         modeling_phi3.CrossEntropyLoss = LigerCrossEntropyLoss
+    if fused_linear_cross_entropy:
+        modeling_phi3.Phi3ForCausalLM.forward = phi3_lce_forward
 
 
 # Model type corresponds to the keys defined in transformers/models/auto/modeling_auto.py
 MODEL_TYPE_TO_APPLY_LIGER_FN = {
     "gemma": apply_liger_kernel_to_gemma,
+    "gemma2": apply_liger_kernel_to_gemma2,
     "llama": apply_liger_kernel_to_llama,
     "mistral": apply_liger_kernel_to_mistral,
     "mixtral": apply_liger_kernel_to_mixtral,

test/convergence/test_mini_models_no_logits.py

@@ -14,13 +14,15 @@
 from transformers.models.gemma2 import Gemma2Config, Gemma2ForCausalLM
 from transformers.models.llama import LlamaConfig, LlamaForCausalLM
 from transformers.models.mistral import MistralConfig, MistralForCausalLM
+from transformers.models.phi3 import Phi3Config, Phi3ForCausalLM
 from transformers.models.qwen2 import Qwen2Config, Qwen2ForCausalLM
 
 from liger_kernel.transformers import (
     apply_liger_kernel_to_gemma,
     apply_liger_kernel_to_gemma2,
     apply_liger_kernel_to_llama,
     apply_liger_kernel_to_mistral,
+    apply_liger_kernel_to_phi3,
     apply_liger_kernel_to_qwen2,
 )
 
@@ -85,6 +87,30 @@
             attn_implementation="sdpa",  # default value, pytorch native attention
         ),
     ),
+    "mini_phi3": MiniModelConfig(
+        liger_kernel_patch_func=apply_liger_kernel_to_phi3,
+        model_class=Phi3ForCausalLM,
+        mini_model_config=Phi3Config(
+            attention_dropout=0.0,
+            bos_token_id=1,
+            eos_token_id=2,  # 32000
+            hidden_act="silu",
+            hidden_size=896,  # 3072
+            initializer_range=0.02,
+            intermediate_size=4864,  # 8192
+            max_position_embeddings=4096,
+            num_attention_heads=8,  # 32
+            num_hidden_layers=4,  # 32
+            num_key_value_heads=None,  # defaults to num_attention_heads
+            rms_norm_eps=1e-5,
+            rope_theta=10000.0,
+            sliding_window=None,
+            tie_word_embeddings=False,
+            use_cache=True,
+            vocab_size=32064,
+            attn_implementation="eager",
+        ),
+    ),
     "mini_mistral": MiniModelConfig(
         liger_kernel_patch_func=apply_liger_kernel_to_mistral,
         model_class=MistralForCausalLM,
@@ -268,6 +294,8 @@ def run_mini_model(
         ("mini_llama3", 32, 1e-4, torch.bfloat16, 5e-3, 1e-5, 1e-1, 1e-5, 1e-2, 1e-5),
         ("mini_qwen2", 32, 1e-4, torch.float32, 1e-8, 1e-5, 5e-3, 1e-5, 5e-3, 1e-5),
         ("mini_qwen2", 32, 1e-4, torch.bfloat16, 1e-8, 1e-5, 1e-1, 1e-5, 1e-2, 1e-5),
+        ("mini_phi3", 32, 1e-4, torch.float32, 1e-8, 1e-5, 5e-3, 1e-5, 5e-3, 1e-5),
+        ("mini_phi3", 32, 1e-4, torch.bfloat16, 1e-8, 1e-5, 1e-1, 1e-5, 1e-2, 1e-5),
         ("mini_mistral", 32, 1e-4, torch.float32, 1e-8, 1e-5, 5e-3, 1e-5, 5e-3, 1e-5),
         ("mini_mistral", 32, 1e-4, torch.bfloat16, 1e-8, 1e-5, 1e-1, 1e-5, 1e-2, 1e-5),
         # Gemma 1.1 and 2 has more tolerance because currently, the kernel is not a perfect match (casts are not done the same way)

test/transformers/test_monkey_patch.py

@@ -16,6 +16,7 @@ def test_import_from_root():
         from liger_kernel.transformers import (  # noqa: F401
             AutoLigerKernelForCausalLM,
             apply_liger_kernel_to_gemma,
+            apply_liger_kernel_to_gemma2,
             apply_liger_kernel_to_llama,
             apply_liger_kernel_to_mistral,
             apply_liger_kernel_to_mixtral,