changes

config.json

@@ -1,31 +1,27 @@
 {
-  "_name_or_path": "microsoft/phi-1_5",
-  "activation_function": "gelu_new",
   "architectures": [
     "PhiForCausalLM"
   ],
-  "attn_pdrop": 0.0,
+  "bos_token_id": 1,
-  "auto_map": {
+  "eos_token_id": 2,
-    "AutoConfig": "configuration_phi.PhiConfig",
+  "hidden_act": "gelu_new",
-    "AutoModelForCausalLM": "modeling_phi.PhiForCausalLM"
+  "hidden_size": 2048,
-  },
-  "embd_pdrop": 0.0,
-  "flash_attn": false,
-  "flash_rotary": false,
-  "fused_dense": false,
   "initializer_range": 0.02,
-  "layer_norm_epsilon": 1e-05,
+  "intermediate_size": 8192,
+  "max_position_embeddings": 2048,
   "model_type": "phi",
-  "n_embd": 2048,
+  "num_attention_heads": 32,
-  "n_head": 32,
+  "num_hidden_layers": 24,
-  "n_head_kv": null,
+  "pretraining_tp": 1,
-  "n_inner": null,
-  "n_layer": 24,
-  "n_positions": 2048,
   "resid_pdrop": 0.0,
-  "rotary_dim": 32,
+  "embd_pdrop": 0.0,
+  "layer_norm_eps": 1e-05,
+  "rope_scaling": null,
+  "rope_theta": 10000.0,
+  "partial_rotary_factor": 0.5,
+  "qk_layernorm": false,
   "tie_word_embeddings": false,
-  "torch_dtype": "float16",
+  "transformers_version": "4.34.0.dev0",
-  "transformers_version": "4.34.1",
+  "use_cache": true,
   "vocab_size": 51200
-}
+}

configuration_phi.py

@@ -1,62 +1,180 @@
-# Copyright (c) Microsoft Corporation.
+# coding=utf-8
-# Licensed under the MIT license.
+# Copyright 2023 Microsoft and the HuggingFace Inc. team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
 
-import math
+""" Phi model configuration"""
-from typing import Optional
 
-from transformers import PretrainedConfig
+
+from transformers.configuration_utils import PretrainedConfig
+from transformers.utils import logging
+
+
+logger = logging.get_logger(__name__)
+
+PHI_PRETRAINED_CONFIG_ARCHIVE_MAP = {
+    "microsoft/phi-1": "https://huggingface.co/microsoft/phi-1/resolve/main/config.json",
+    "microsoft/phi-1_5": "https://huggingface.co/microsoft/phi-1_5/resolve/main/config.json",
+}
 
 
 class PhiConfig(PretrainedConfig):
-    """Phi configuration."""
+    r"""
+    This is the configuration class to store the configuration of a [`PhiModel`]. It is used to instantiate an Phi
+    model according to the specified arguments, defining the model architecture. Instantiating a configuration with the
+    defaults will yield a similar configuration to that of the Phi
+    [microsoft/phi-1](https://huggingface.co/microsoft/phi-1).
+
+    Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
+    documentation from [`PretrainedConfig`] for more information.
+
+    Args:
+        vocab_size (`int`, *optional*, defaults to 51200):
+            Vocabulary size of the Phi model. Defines the number of different tokens that can be represented by the
+            `inputs_ids` passed when calling [`PhiModel`].
+        hidden_size (`int`, *optional*, defaults to 2048):
+            Dimension of the hidden representations.
+        intermediate_size (`int`, *optional*, defaults to 8192):
+            Dimension of the MLP representations.
+        num_hidden_layers (`int`, *optional*, defaults to 24):
+            Number of hidden layers in the Transformer decoder.
+        num_attention_heads (`int`, *optional*, defaults to 32):
+            Number of attention heads for each attention layer in the Transformer decoder.
+        resid_pdrop (`float`, *optional*, defaults to 0.0):
+            Dropout probability for mlp outputs.
+        embd_pdrop (`int`, *optional*, defaults to 0.0):
+            The dropout ratio for the embeddings.
+        attention_dropout (`float`, *optional*, defaults to 0.0):
+            The dropout ratio after computing the attention scores.
+        hidden_act (`str` or `function`, *optional*, defaults to `"gelu_new"`):
+            The non-linear activation function (function or string) in the decoder.
+        max_position_embeddings (`int`, *optional*, defaults to 2048):
+            The maximum sequence length that this model might ever be used with. Phi-1 and Phi-1.5 supports up to 2048
+            tokens.
+        initializer_range (`float`, *optional*, defaults to 0.02):
+            The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
+        layer_norm_eps (`float`, *optional*, defaults to 1e-05):
+            The epsilon used by the rms normalization layers.
+        use_cache (`bool`, *optional*, defaults to `True`):
+            Whether or not the model should return the last key/values attentions (not used by all models). Only
+            relevant if `config.is_decoder=True`. Whether to tie weight embeddings or not.
+        tie_word_embeddings (`bool`, *optional*, defaults to `False`):
+            Whether to tie weight embeddings
+        rope_theta (`float`, *optional*, defaults to 10000.0):
+            The base period of the RoPE embeddings.
+        rope_scaling (`Dict`, *optional*):
+            Dictionary containing the scaling configuration for the RoPE embeddings. Currently supports two scaling
+            strategies: linear and dynamic. Their scaling factor must be an float greater than 1. The expected format
+            is `{"type": strategy name, "factor": scaling factor}`. When using this flag, don't update
+            `max_position_embeddings` to the expected new maximum. See the following thread for more information on how
+            these scaling strategies behave:
+            https://www.reddit.com/r/LocalPersimmon/comments/14mrgpr/dynamically_scaled_rope_further_increases/. This
+            is an experimental feature, subject to breaking API changes in future versions.
+        partial_rotary_factor (`float`, *optional*, defaults to 0.5):
+            Percentage of the query and keys which will have rotary embedding.
+        qk_layernorm (`bool`, *optional*, defaults to `False`):
+            Whether or not to normalize the Queries and Keys after projecting the hidden states
+        bos_token_id (`int`, *optional*, defaults to 1):
+            Denotes beginning of sequences token id.
+        eos_token_id (`int`, *optional*, defaults to 2):
+            Denotes end of sequences token id.
+
+    Example:
+
+    ```python
+    >>> from transformers import PhiModel, PhiConfig
+
+    >>> # Initializing a Phi-1 style configuration
+    >>> configuration = PhiConfig.from_pretrained("microsoft/phi-1")
+
+    >>> # Initializing a model from the configuration
+    >>> model = PhiModel(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+    ```"""
 
     model_type = "phi"
-    attribute_map = {
+    keys_to_ignore_at_inference = ["past_key_values"]
-        "max_position_embeddings": "n_positions",
-        "hidden_size": "n_embd",
-        "num_attention_heads": "n_head",
-        "num_hidden_layers": "n_layer",
-    }
 
     def __init__(
         self,
-        vocab_size: int = 50304,
+        vocab_size=51200,
-        n_positions: int = 2048,
+        hidden_size=2048,
-        n_embd: int = 1024,
+        intermediate_size=8192,
-        n_layer: int = 20,
+        num_hidden_layers=24,
-        n_inner: Optional[int] = None,
+        num_attention_heads=32,
-        n_head: int = 16,
+        resid_pdrop=0.0,
-        n_head_kv: Optional[int] = None,
+        embd_pdrop=0.0,
-        rotary_dim: Optional[int] = 32,
+        attention_dropout=0.0,
-        activation_function: Optional[str] = "gelu_new",
+        hidden_act="gelu_new",
-        flash_attn: bool = False,
+        max_position_embeddings=2048,
-        flash_rotary: bool = False,
+        initializer_range=0.02,
-        fused_dense: bool = False,
+        layer_norm_eps=1e-5,
-        attn_pdrop: float = 0.0,
+        use_cache=True,
-        embd_pdrop: float = 0.0,
+        tie_word_embeddings=False,
-        resid_pdrop: float = 0.0,
+        rope_theta=10000.0,
-        layer_norm_epsilon: float = 1e-5,
+        rope_scaling=None,
-        initializer_range: float = 0.02,
+        partial_rotary_factor=0.5,
-        tie_word_embeddings: bool = False,
+        qk_layernorm=False,
-        pad_vocab_size_multiple: int = 64,
+        bos_token_id=1,
-        **kwargs
+        eos_token_id=2,
-    ) -> None:
+        **kwargs,
-        self.vocab_size = int(math.ceil(vocab_size / pad_vocab_size_multiple) * pad_vocab_size_multiple)
+    ):
-        self.n_positions = n_positions
+        self.vocab_size = vocab_size
-        self.n_embd = n_embd
+        self.hidden_size = hidden_size
-        self.n_layer = n_layer
+        self.intermediate_size = intermediate_size
-        self.n_inner = n_inner
+        self.num_hidden_layers = num_hidden_layers
-        self.n_head = n_head
+        self.num_attention_heads = num_attention_heads
-        self.n_head_kv = n_head_kv
-        self.rotary_dim = min(rotary_dim, n_embd // n_head)
-        self.activation_function = activation_function
-        self.flash_attn = flash_attn
-        self.flash_rotary = flash_rotary
-        self.fused_dense = fused_dense
-        self.attn_pdrop = attn_pdrop
-        self.embd_pdrop = embd_pdrop
         self.resid_pdrop = resid_pdrop
-        self.layer_norm_epsilon = layer_norm_epsilon
+        self.embd_pdrop = embd_pdrop
+        self.attention_dropout = attention_dropout
+        self.hidden_act = hidden_act
+        self.max_position_embeddings = max_position_embeddings
         self.initializer_range = initializer_range
+        self.layer_norm_eps = layer_norm_eps
+        self.use_cache = use_cache
+        self.rope_theta = rope_theta
+        self.rope_scaling = rope_scaling
+        self.partial_rotary_factor = partial_rotary_factor
+        self.qk_layernorm = qk_layernorm
+        self._rope_scaling_validation()
 
-        super().__init__(tie_word_embeddings=tie_word_embeddings, **kwargs)
+        super().__init__(
+            bos_token_id=bos_token_id,
+            eos_token_id=eos_token_id,
+            tie_word_embeddings=tie_word_embeddings,
+            **kwargs,
+        )
+
+    # Copied from transformers.models.llama.configuration_llama.LlamaConfig._rope_scaling_validation
+    def _rope_scaling_validation(self):
+        """
+        Validate the `rope_scaling` configuration.
+        """
+        if self.rope_scaling is None:
+            return
+
+        if not isinstance(self.rope_scaling, dict) or len(self.rope_scaling) != 2:
+            raise ValueError(
+                "`rope_scaling` must be a dictionary with with two fields, `type` and `factor`, "
+                f"got {self.rope_scaling}"
+            )
+        rope_scaling_type = self.rope_scaling.get("type", None)
+        rope_scaling_factor = self.rope_scaling.get("factor", None)
+        if rope_scaling_type is None or rope_scaling_type not in ["linear", "dynamic"]:
+            raise ValueError(
+                f"`rope_scaling`'s type field must be one of ['linear', 'dynamic'], got {rope_scaling_type}"
+            )
+        if rope_scaling_factor is None or not isinstance(rope_scaling_factor, float) or rope_scaling_factor <= 1.0:
+            raise ValueError(f"`rope_scaling`'s factor field must be a float > 1, got {rope_scaling_factor}")