# Copyright 2022 Cerebras Systems.
#
# 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 logging
import re
from typing import Tuple
import torch
from cerebras.modelzoo.tools.checkpoint_converters.base_converter import (
BaseCheckpointConverter_CS_CS,
BaseCheckpointConverter_HF_CS,
BaseConfigConverter,
BaseConfigConverter_CS_CS,
BaseConfigConverter_HF_CS,
ConfigConversionError,
ConversionRule,
EquivalentSubkey,
FormatVersions,
)
from cerebras.modelzoo.tools.checkpoint_converters.helper import (
Build_HF_CS_Converter_WithOptionalModel,
convert_use_rms_layer_norm_helper,
maybe_tie_lm_head,
tie_none_weights,
transpose_key_if_2D,
)
#########################################################
# GPT2 HF <> CS17
#########################################################
[docs]class Converter_GPT2_Attention_HF_CS17(BaseCheckpointConverter_HF_CS):
def __init__(self, generate_hf_biases=True):
super().__init__()
self.generate_hf_biases = generate_hf_biases
self.rules = [
ConversionRule(
[
EquivalentSubkey("c_proj", "proj_output_dense_layer"),
r"\.(?:weight|bias)",
],
action=transpose_key_if_2D,
),
ConversionRule(
[
EquivalentSubkey("c_attn", "proj_q_dense_layer"),
r"\.(?:weight|bias)",
],
action=self.c_attn_converter,
),
ConversionRule(
[
EquivalentSubkey("q_attn", "proj_q_dense_layer"),
r"\.(?:weight|bias)",
],
action=self.assert_already_converted,
),
ConversionRule(
[
EquivalentSubkey("c_attn", "proj_k_dense_layer"),
r"\.(?:weight|bias)",
],
action=self.assert_already_converted,
),
ConversionRule(
[
EquivalentSubkey("c_attn", "proj_v_dense_layer"),
r"\.(?:weight|bias)",
],
action=self.assert_already_converted,
),
]
@staticmethod
def formats() -> Tuple[FormatVersions, FormatVersions]:
return (FormatVersions("hf"), FormatVersions("cs-1.7"))
@staticmethod
def get_config_converter_class() -> BaseConfigConverter:
return None
def c_attn_converter(
self,
old_key,
new_key,
old_state_dict,
new_state_dict,
from_index,
action_fn_args,
):
if from_index == 0:
self.c_attn_converter_hf_to_cs17(
old_key, new_key, old_state_dict, new_state_dict, action_fn_args
)
else:
self.c_attn_converter_cs17_to_hf(
old_key, new_key, old_state_dict, new_state_dict, action_fn_args
)
def c_attn_converter_hf_to_cs17(
self, old_key, new_key, old_state_dict, new_state_dict, action_fn_args
):
# HF represents Q, K, and V in a packed format. We need to unpack the
# weight and bias tensor for CS 1.7 format.
q_key = new_key
k_key = re.sub(r"\.proj_q_dense_layer\.", ".proj_k_dense_layer.", q_key)
v_key = re.sub(r"\.proj_q_dense_layer\.", ".proj_v_dense_layer.", q_key)
if new_key.endswith(".bias"):
assert len(old_state_dict[old_key].shape) == 1
packed_dim = old_state_dict[old_key].shape[0]
embed_dim = packed_dim // 3
assert 3 * embed_dim == packed_dim, (
f"Invalid tensor shape {old_state_dict[old_key].shape} at {old_key}. Bias should "
f"be divisible by 3 since Q, K, and V are packed."
)
(
new_state_dict[q_key],
new_state_dict[k_key],
new_state_dict[v_key],
) = torch.chunk(old_state_dict[old_key], 3, dim=0)
elif new_key.endswith(".weight"):
embed_dim, packed_dim = old_state_dict[old_key].shape
assert 3 * embed_dim == packed_dim, (
f"Invalid tensor shape {old_state_dict[old_key].shape} at {old_key}. The second "
f"dimension should be 3x the first dimension (embed_dim) since Q, K, and V are "
f"packed."
)
(
new_state_dict[q_key],
new_state_dict[k_key],
new_state_dict[v_key],
) = torch.chunk(
torch.transpose(old_state_dict[old_key], 0, 1), 3, dim=0
)
else:
raise ValueError("Invalid key after conversion: {}".format(new_key))
def c_attn_converter_cs17_to_hf(
self,
old_key,
new_key,
old_state_dict,
new_state_dict,
action_fn_args,
):
# HF represents Q, K, and V in a packed format. It also contains
# special ".bias" and ".masked_bias" register buffers that need to be
# initialized
q_key = old_key
k_key = re.sub(r"\.proj_q_dense_layer\.", ".proj_k_dense_layer.", q_key)
v_key = re.sub(r"\.proj_q_dense_layer\.", ".proj_v_dense_layer.", q_key)
assert (
k_key in old_state_dict
), "Expected the following key to exist! {}".format(k_key)
assert (
v_key in old_state_dict
), "Expected the following key to exist! {}".format(v_key)
new_state_dict[new_key] = torch.cat(
(
old_state_dict[q_key],
old_state_dict[k_key],
old_state_dict[v_key],
),
dim=0,
)
# Need to transpose to convert from Linear.weight -> Conv1D.weight
if len(new_state_dict[new_key].shape) == 2:
new_state_dict[new_key] = torch.transpose(
new_state_dict[new_key], 0, 1
)
if new_key.endswith(".bias") and self.generate_hf_biases:
max_position_embeddings = action_fn_args["configs"][1]["model"][
"max_position_embeddings"
]
attn_bias_key = re.sub(r"\.c_attn\.", ".", new_key)
new_state_dict[attn_bias_key] = torch.tril(
torch.ones(
(max_position_embeddings, max_position_embeddings),
dtype=torch.uint8,
)
).view(1, 1, max_position_embeddings, max_position_embeddings)
masked_bias_key = re.sub(r"\.c_attn\.", ".masked_", new_key)
new_state_dict[masked_bias_key] = torch.tensor(-1e4)
def assert_already_converted(
self,
old_key,
new_key,
old_state_dict,
new_state_dict,
from_index,
action_fn_args,
):
if from_index == 0:
# We should never hit this case as this key should have been matched
# already
assert False, "Invalid key: {}".format(old_key)
else:
# When we convert from CS -> HF, the proj_q_dense_layer should also handle
# conversion of proj_k_dense_layer and proj_v_dense_layer since HF
# represents these three layers in a packed format. We simply need
# to test that the key containing the packed format has already
# been converted.
assert (
new_key in new_state_dict
), "Key should've been already converted: {} -> {}".format(
old_key, new_key
)
[docs]class Converter_GPT2Model_HF_CS17(BaseCheckpointConverter_HF_CS):
def __init__(self):
super().__init__()
self.rules = [
ConversionRule(
[
EquivalentSubkey("wte", "embedding_layer.word_embeddings"),
r"\.(?:weight|bias)",
],
action=self.replaceKey,
),
ConversionRule(
[
EquivalentSubkey(
"wpe", "embedding_layer.position_embeddings"
),
r"\.(?:weight|bias)",
],
action=self.replaceKey,
),
ConversionRule(
[
EquivalentSubkey("h", "transformer_decoder.layers"),
r"\.\d+\.",
EquivalentSubkey("attn.", "self_attn."),
self.attention_converter_class(),
],
action=None,
),
ConversionRule(
[
EquivalentSubkey("h", "transformer_decoder.layers"),
r"\.\d+\.",
EquivalentSubkey("ln_1", "norm1"),
r"\.(?:weight|bias)",
],
action=self.replaceKey,
),
ConversionRule(
[
EquivalentSubkey("h", "transformer_decoder.layers"),
r"\.\d+\.",
EquivalentSubkey("ln_2", "norm3"),
r"\.(?:weight|bias)",
],
action=self.replaceKey,
),
ConversionRule(
[
EquivalentSubkey("h", "transformer_decoder.layers"),
r"\.\d+\.",
EquivalentSubkey("mlp.c_fc", "ffn.ffn.0.linear_layer"),
r"\.(?:weight|bias)",
],
action=self.ffn_converter(),
),
ConversionRule(
[
EquivalentSubkey("h", "transformer_decoder.layers"),
r"\.\d+\.",
EquivalentSubkey("mlp.c_proj", "ffn.ffn.1.linear_layer"),
r"\.(?:weight|bias)",
],
action=self.ffn_converter(),
),
ConversionRule(
[
EquivalentSubkey("ln_f", "transformer_decoder.norm"),
r"\.(?:weight|bias)",
],
action=self.replace_final_norm,
),
ConversionRule([r"lm_head\.(?:weight|bias)"], exists="right"),
ConversionRule([r"ln_f\.(?:weight|bias)"], exists="right"),
ConversionRule(
[
r"h\.\d+\.attn\.(?:masked_bias|bias)",
],
exists="left",
),
]
def attention_converter_class(self):
# Allows other checkpoint converters to inherit from
# this main converter but can overide this function with
# different types of attention converters (i.e. MQA)
return Converter_GPT2_Attention_HF_CS17()
def ffn_converter(self):
# similar to above, allows overriding method for other models
# that use mostly GPT-2, but with slight changes
return transpose_key_if_2D
def replace_final_norm(
self,
old_key,
new_key,
old_state_dict,
new_state_dict,
from_index,
action_fn_args,
):
new_state_dict[new_key] = old_state_dict[old_key]
# CS 1.7 has both "ln_f" and "transformer_decoder.norm"
# we need to copy the original ("ln_f") too:
if from_index == 0:
ln_f_key = re.sub(r"transformer_decoder\.norm\.", "ln_f.", new_key)
new_state_dict[ln_f_key] = old_state_dict[old_key]
def pre_model_convert(
self,
old_state_dict,
new_state_dict,
configs,
converter_indices,
drop_unmatched_keys,
):
if converter_indices.direction == 0:
logging.warning(
"{} GPT2 has a language model head (lm_head) "
"while {} GPT2Model does not. Initializing lm_head to default.".format(
self.formats()[1], self.formats()[0]
)
)
# Manually tie weights
if (
converter_indices.direction == 1
and configs[1]["model"]["share_embedding_weights"]
):
if (
old_state_dict.get("embedding_layer.word_embeddings.weight", 0)
is None
):
old_state_dict["embedding_layer.word_embeddings.weight"] = (
old_state_dict["lm_head.weight"]
)
def post_model_convert(
self,
old_state_dict,
new_state_dict,
configs,
converter_indices,
drop_unmatched_keys,
key_prefix="",
):
if converter_indices.direction == 0:
# We are converting from HF GPT2Model (which is headless) -> CS GPT2LMHeadModel
# We need to create 'lm_head' and init to default values
hf_config = configs[0]
cs_config = configs[1]
use_bias_in_output = cs_config["model"].get(
"use_bias_in_output", False
)
vocab_size = cs_config["model"]["vocab_size"]
embed_dim = cs_config["model"]["hidden_size"]
if hf_config["tie_word_embeddings"]:
lm_head_weight = old_state_dict['wte.weight']
else:
lm_head_weight = torch.zeros((vocab_size, embed_dim))
lm_head_weight.normal_(mean=0.0, std=0.02)
new_state_dict[key_prefix + "lm_head.weight"] = lm_head_weight
if use_bias_in_output:
lm_head_bias = torch.zeros(vocab_size)
new_state_dict[key_prefix + "lm_head.bias"] = lm_head_bias
super().post_model_convert(
old_state_dict,
new_state_dict,
configs,
converter_indices,
drop_unmatched_keys,
key_prefix=key_prefix,
)
@staticmethod
def formats() -> Tuple[FormatVersions, FormatVersions]:
return (FormatVersions("hf"), FormatVersions("cs-1.7"))
@classmethod
def converter_note(cls) -> str:
return (
"{} GPT2Model <-> {} GPT2LMHeadModel\n"
"The HF model doesn't contain a language model head while the CS "
"one does. When converting to CS, the exported checkpoint will "
"contain a language model head initialized to default random "
"values. When converting to HF, the language model head will be "
"dropped."
).format(cls.formats()[0], cls.formats()[1])
@staticmethod
def get_config_converter_class() -> BaseConfigConverter:
return ConfigConverter_GPT2Model_HF_CS17
class Converter_GPT2LMHeadModel_HF_CS17(BaseCheckpointConverter_HF_CS):
def __init__(self):
super().__init__()
self.rules = [
ConversionRule(
["lm_head\.(?:weight|bias)"],
action=self.replaceKey,
),
ConversionRule(
[
EquivalentSubkey("transformer.", ""),
Converter_GPT2Model_HF_CS17(),
],
action=None,
),
]
def pre_model_convert(
self,
old_state_dict,
new_state_dict,
configs,
converter_indices,
drop_unmatched_keys,
):
# Manually tie weights
if (
converter_indices.direction == 1
and configs[1]["model"]["share_embedding_weights"]
):
if (
old_state_dict.get("embedding_layer.word_embeddings.weight", 0)
is None
):
old_state_dict["embedding_layer.word_embeddings.weight"] = (
old_state_dict["lm_head.weight"]
)
@staticmethod
def formats() -> Tuple[FormatVersions, FormatVersions]:
return (FormatVersions("hf"), FormatVersions("cs-1.7"))
@classmethod
def converter_note(cls) -> str:
return "{} GPT2LMHeadModel <-> {} GPT2LMHeadModel".format(
cls.formats()[0], cls.formats()[1]
)
@staticmethod
def get_config_converter_class() -> BaseConfigConverter:
return ConfigConverter_GPT2Model_HF_CS17
[docs]class ConfigConverter_GPT2Model_HF_CS17(BaseConfigConverter_HF_CS):
def __init__(self):
super().__init__()
self.rules = [
ConversionRule(
["model_type"],
action=BaseConfigConverter.assert_factory_fn(0, "gpt2"),
),
# Embedding
ConversionRule(["vocab_size"], action=self.replaceKey),
ConversionRule(
["position_embedding_type"],
exists="right",
action=BaseConfigConverter.assert_factory_fn(1, "learned"),
),
ConversionRule(
["use_position_embedding"],
exists="right",
action=BaseConfigConverter.assert_factory_fn(1, True),
),
ConversionRule(
[EquivalentSubkey("embd_pdrop", "embedding_dropout_rate")],
action=self.replaceKey,
),
ConversionRule(
[
EquivalentSubkey(
"tie_word_embeddings", "share_embedding_weights"
)
],
action=self.replaceKey,
),
ConversionRule(
["embedding_layer_norm"],
action=BaseConfigConverter.assert_factory_fn(1, False),
),
# Decoder Block
ConversionRule(
[EquivalentSubkey("n_embd", "hidden_size")],
action=self.replaceKey,
),
ConversionRule(
[EquivalentSubkey("n_head", "num_heads")],
action=self.replaceKey,
),
ConversionRule(
[EquivalentSubkey("n_layer", "num_hidden_layers")],
action=self.replaceKey,
),
ConversionRule(
[EquivalentSubkey("n_positions", "max_position_embeddings")],
action=self.replaceKey,
),
ConversionRule(
[EquivalentSubkey("scale_attn_weights", "attention_type")],
action=self.convert_attention_type,
),
ConversionRule(
["use_projection_bias_in_attention"],
action=BaseConfigConverter.assert_factory_fn(1, True),
),
ConversionRule(
["use_ffn_bias_in_attention"],
exists="right",
action=BaseConfigConverter.assert_factory_fn(1, True),
),
ConversionRule(
["use_ffn_bias"],
exists="right",
action=BaseConfigConverter.assert_factory_fn(1, True),
),
ConversionRule(
[EquivalentSubkey("n_inner", "filter_size")],
action=self.replaceKey,
),
ConversionRule(
[EquivalentSubkey("activation_function", "nonlinearity")],
action=self.replaceKey,
),
ConversionRule(
[EquivalentSubkey("attn_pdrop", "attention_dropout_rate")],
action=self.replaceKey,
),
ConversionRule(
[EquivalentSubkey("resid_pdrop", "dropout_rate")],
action=self.replaceKey,
),
ConversionRule(["rotary_dim"], action=self.replaceKey),
ConversionRule(
["layer_norm_epsilon"],
action=self.replaceKey,
),
ConversionRule(
["use_bias_in_output"],
action=BaseConfigConverter.assert_factory_fn(1, False),
),
ConversionRule(["initializer_range"], action=self.replaceKey),
ConversionRule(
["fixed_sparse_attention"],
action=BaseConfigConverter.assert_factory_fn(1, None),
),
ConversionRule(
["norm_first"],
action=BaseConfigConverter.assert_factory_fn(1, True),
),
ConversionRule(
["use_ff_layer1_dropout"],
action=BaseConfigConverter.assert_factory_fn(1, False),
),
ConversionRule(
["scale_attn_by_inverse_layer_idx"],
action=BaseConfigConverter.assert_factory_fn(0, False),
),
ConversionRule(
["reorder_and_upcast_attn"],
action=BaseConfigConverter.assert_factory_fn(0, False),
),
]
self.pre_convert_defaults[0].update(
{
"tie_word_embeddings": True,
}
)
self.pre_convert_defaults[1].update(
{
"share_embedding_weights": True,
},
)
self.post_convert_defaults[0].update({"model_type": "gpt2"})
self.post_convert_defaults[1].update({"use_bias_in_output": False})
def convert_attention_type(
self,
old_key,
new_key,
old_state_dict,
new_state_dict,
from_index,
action_fn_args,
):
if from_index == 0:
new_state_dict[new_key] = (
"scaled_dot_product"
if old_state_dict[old_key]
else "dot_product"
)
else:
if (
old_state_dict[old_key] != "scaled_dot_product"
and old_state_dict[old_key] != "dot_product"
):
raise ConfigConversionError(
"Can't convert config with {}={}. Only {} is supported.".format(
old_key,
old_state_dict[old_key],
"scaled_dot_product and dot_product",
)
)
new_state_dict[new_key] = old_state_dict[old_key].startswith(
"scaled_"
)
def pre_config_convert(
self,
model,
config,
converter_indices,
):
config = super().pre_config_convert(model, config, converter_indices)
if converter_indices.direction == 0:
if "n_inner" not in config or config["n_inner"] is None:
config["n_inner"] = 4 * config["n_embd"]
else:
if "embedding_dropout_rate" not in config:
config["embedding_dropout_rate"] = config["dropout_rate"]
if "attention_dropout_rate" not in config:
config["attention_dropout_rate"] = config["dropout_rate"]
return config
@staticmethod
def formats() -> Tuple[FormatVersions, FormatVersions]:
return (FormatVersions("hf"), FormatVersions("cs-1.7"))
#########################################################
# GPT2 HF <> CS18, CS19
#########################################################
[docs]class ConfigConverter_GPT2Model_HF_CS18(ConfigConverter_GPT2Model_HF_CS17):
@staticmethod
def formats() -> Tuple[FormatVersions, FormatVersions]:
return (FormatVersions("hf"), FormatVersions("cs-1.8", "cs-1.9"))
def supports_mup_conversion(self):
return True
Converter_GPT2Model_HF_CS18 = Build_HF_CS_Converter_WithOptionalModel(
"Converter_GPT2Model_HF_CS18",
Converter_GPT2Model_HF_CS17,
derived_class=Converter_GPT2Model_HF_CS17,
config_converter_class=ConfigConverter_GPT2Model_HF_CS18,
formats=(FormatVersions("hf"), FormatVersions("cs-1.8", "cs-1.9")),
converter_note_fn=lambda cls: (
"{} GPT2Model <-> {} GPT2LMHeadModel\n"
"The HF model doesn't contain a language model head while the CS "
"one does. When converting to CS, the exported checkpoint will "
"contain a language model head initialized to default random "
"values. When converting to HF, the language model head will be "
"dropped."
).format(cls.formats()[0], cls.formats()[1]),
)
[docs]class Converter_GPT2LMHeadModel_HF_CS17(BaseCheckpointConverter_HF_CS):
def __init__(self):
super().__init__()
self.rules = [
ConversionRule(
[r"lm_head\.(?:weight|bias)"],
action=self.replaceKey,
),
ConversionRule(
[
EquivalentSubkey("transformer.", ""),
Converter_GPT2Model_HF_CS17(),
],
action=None,
),
]
def pre_model_convert(
self,
old_state_dict,
new_state_dict,
configs,
converter_indices,
drop_unmatched_keys,
):
# Manually tie weights
if (
converter_indices.direction == 1
and configs[1]["model"]["share_embedding_weights"]
):
if (
old_state_dict.get("embedding_layer.word_embeddings.weight", 0)
is None
):
old_state_dict["embedding_layer.word_embeddings.weight"] = (
old_state_dict["lm_head.weight"]
)
@staticmethod
def formats() -> Tuple[FormatVersions, FormatVersions]:
return (FormatVersions("hf"), FormatVersions("cs-1.7"))
@classmethod
def converter_note(cls) -> str:
return "{} GPT2LMHeadModel <-> {} GPT2LMHeadModel".format(
cls.formats()[0], cls.formats()[1]
)
@staticmethod
def get_config_converter_class() -> BaseConfigConverter:
return ConfigConverter_GPT2Model_HF_CS17
[docs]class Converter_GPT2LMHeadModel_HF_CS18(BaseCheckpointConverter_HF_CS):
def __init__(self):
super().__init__()
self.rules = [
# Catch checkpoints from Pytorch 2.0 API
ConversionRule(
[
Converter_GPT2LMHeadModel_HF_CS17(),
],
action=None,
),
# Catch checkpoints from 1.7/1.8
ConversionRule(
[
EquivalentSubkey("", "model."),
Converter_GPT2LMHeadModel_HF_CS17(),
],
action=None,
),
]
def post_model_convert(
self,
old_state_dict,
new_state_dict,
configs,
converter_indices,
drop_unmatched_keys,
key_prefix="",
):
if converter_indices.direction == 0:
lm_head_weight_key = key_prefix + "lm_head.weight"
embed_key = key_prefix + "transformer.wte.weight"
if lm_head_weight_key not in new_state_dict:
new_state_dict[lm_head_weight_key] = old_state_dict[embed_key]
super().post_model_convert(
old_state_dict,
new_state_dict,
configs,
converter_indices,
drop_unmatched_keys,
key_prefix=key_prefix,
)
@staticmethod
def formats() -> Tuple[FormatVersions, FormatVersions]:
return (FormatVersions("hf"), FormatVersions("cs-1.8", "cs-1.9"))
@classmethod
def converter_note(cls) -> str:
return "{} GPT2LMHeadModel <-> {} GPT2LMHeadModel".format(
cls.formats()[0], cls.formats()[1]
)
@staticmethod
def get_config_converter_class() -> BaseConfigConverter:
return ConfigConverter_GPT2Model_HF_CS18
def supports_mup_conversion(self) -> bool:
return True
###########################################################
# In CS 2.0, we changed introduced norm_type in the config.
# CS 1.8, CS 1.9 <> CS 2.0, and HF <> CS 2.0 converters:
###########################################################
[docs]class Converter_GPT2LMHeadModel_CS18_CS20(BaseCheckpointConverter_CS_CS):
def __init__(self):
super().__init__()
# Model didn't change between 1.8/1.9 and 2.0. Copy all keys.
self.rules = [
ConversionRule(
[
"(?:model.|)",
EquivalentSubkey(
"lm_head", "embedding_layer.word_embeddings"
),
"\.weight",
],
action=maybe_tie_lm_head,
),
ConversionRule(
[
"(?:model.|)",
EquivalentSubkey(
"embedding_layer.word_embeddings",
"lm_head",
),
"\.weight",
],
action=maybe_tie_lm_head,
),
ConversionRule(
[
"(?:model.|)",
EquivalentSubkey("transformer_decoder.norm", "ln_f"),
"\.(?:weight|bias)",
],
action=tie_none_weights,
),
ConversionRule(
[
"(?:model.|)",
EquivalentSubkey("ln_f", "transformer_decoder.norm"),
"\.(?:weight|bias)",
],
action=tie_none_weights,
),
ConversionRule([".*"], action=self.replaceKey),
]
@classmethod
def converter_note(cls) -> str:
return "GPT2LMHeadModel class"
@staticmethod
def formats() -> Tuple[FormatVersions, FormatVersions]:
return (FormatVersions("cs-1.8", "cs-1.9"), FormatVersions("cs-2.0"))
@staticmethod
def get_config_converter_class() -> BaseConfigConverter:
return ConfigConverter_GPT2Model_CS18_CS20
[docs]class ConfigConverter_GPT2Model_CS18_CS20(BaseConfigConverter_CS_CS):
def __init__(self):
super().__init__()
# Only difference between 1.8/1.9 and 2.0 is introduction of norm_type
self.rules = [
ConversionRule(
[EquivalentSubkey("use_rms_norm", "norm_type")],
action=self.convert_use_rms_layer_norm,
),
ConversionRule([".*"], action=self.replaceKey),
]
self.pre_convert_defaults[0]["use_rms_norm"] = False
self.pre_convert_defaults[1]["norm_type"] = "layernorm"
def convert_use_rms_layer_norm(self, *args):
convert_use_rms_layer_norm_helper(self, *args)
@staticmethod
def formats() -> Tuple[FormatVersions, FormatVersions]:
return (FormatVersions("cs-1.8", "cs-1.9"), FormatVersions("cs-2.0"))
[docs]class Converter_GPT2Model_HF_CS20(Converter_GPT2Model_HF_CS18):
@staticmethod
def formats() -> Tuple[FormatVersions, FormatVersions]:
return (FormatVersions("hf"), FormatVersions("cs-2.0"))
@staticmethod
def get_config_converter_class() -> BaseConfigConverter:
return ConfigConverter_GPT2Model_HF_CS20
[docs]class Converter_GPT2LMHeadModel_HF_CS20(Converter_GPT2LMHeadModel_HF_CS18):
@staticmethod
def formats() -> Tuple[FormatVersions, FormatVersions]:
return (FormatVersions("hf"), FormatVersions("cs-2.0"))
@staticmethod
def get_config_converter_class() -> BaseConfigConverter:
return ConfigConverter_GPT2Model_HF_CS20
[docs]class ConfigConverter_GPT2Model_HF_CS20(ConfigConverter_GPT2Model_HF_CS18):
def __init__(self):
super().__init__()
self.rules = [
ConversionRule(
["norm_type"],
action=BaseConfigConverter.assert_factory_fn(1, "layernorm"),
),
*self.rules,
]
@staticmethod
def formats() -> Tuple[FormatVersions, FormatVersions]:
return (FormatVersions("hf"), FormatVersions("cs-2.0"))
###########################################################
# In CS 2.1, we refactored the embedding layer.
# CS 2.0 <> CS 2.1, and HF <> CS 2.1 converters:
###########################################################
[docs]class Converter_GPT2LMHeadModel_CS20_CS21(BaseCheckpointConverter_CS_CS):
def __init__(self):
super().__init__()
self.rules = [
# Refactored embeddings:
ConversionRule(
[
"(?:model\.|)",
EquivalentSubkey(
"embedding_layer.position_embeddings.weight",
"embedding_layer.position_embeddings.embed.weight",
),
],
action=self.replaceKey,
),
ConversionRule(
[
"(?:model\.|)",
"embedding_layer\.",
EquivalentSubkey(
"position_embeddings",
"position_embeddings.fpe",
),
],
action=self.replaceKey,
),
ConversionRule(
[
"(?:model\.|)",
EquivalentSubkey(
"relative_pe_helper.relative_attention_bias",
"embedding_layer.position_embed_helper.relative_attention_bias",
),
"\.(?:weight|bias)",
],
action=self.replaceKey,
),
ConversionRule(
[
"(?:model\.|)",
EquivalentSubkey(
"relative_pe_helper.slopes",
"embedding_layer.position_embed_helper.slopes",
),
],
action=self.replaceKey,
),
# Copy everything else
ConversionRule([".*"], action=self.replaceKey),
]
@classmethod
def converter_note(cls) -> str:
return "GPT2LMHeadModel class"
@staticmethod
def formats() -> Tuple[FormatVersions, FormatVersions]:
return (FormatVersions("cs-2.0"), FormatVersions("cs-2.1"))
@staticmethod
def get_config_converter_class() -> BaseConfigConverter:
return ConfigConverter_GPT2Model_CS20_CS21
[docs]class ConfigConverter_GPT2Model_CS20_CS21(BaseConfigConverter_CS_CS):
def __init__(self):
super().__init__()
self.rules = [
ConversionRule([".*"], action=self.replaceKey),
]
@staticmethod
def formats() -> Tuple[FormatVersions, FormatVersions]:
return (FormatVersions("cs-2.0"), FormatVersions("cs-2.1"))
[docs]class ConfigConverter_GPT2Model_HF_CS21(ConfigConverter_GPT2Model_HF_CS20):
"CS 2.1 config is the same as CS 2.0."
@staticmethod
def formats() -> Tuple[FormatVersions, FormatVersions]:
return (
FormatVersions("hf"),
FormatVersions("cs-2.1", "cs-2.2", "cs-2.3", "cs-2.4"),
)
[docs]class Converter_GPT2Model_WithoutOptionalModel_HF_CS21(
Converter_GPT2Model_HF_CS17
):
def __init__(self):
super().__init__()
self.rules = [
ConversionRule(
[
EquivalentSubkey(
"wpe", "embedding_layer.position_embeddings.embed"
),
"\.(?:weight|bias)",
],
action=self.replaceKey,
),
*self.rules,
]
def supports_mup_conversion(self) -> bool:
return True
Converter_GPT2Model_HF_CS21 = Build_HF_CS_Converter_WithOptionalModel(
"Converter_GPT2Model_HF_CS21",
Converter_GPT2Model_WithoutOptionalModel_HF_CS21,
derived_class=Converter_GPT2Model_WithoutOptionalModel_HF_CS21,
config_converter_class=ConfigConverter_GPT2Model_HF_CS21,
formats=(
FormatVersions("hf"),
FormatVersions("cs-2.1", "cs-2.2", "cs-2.3", "cs-2.4"),
),
)
[docs]class Converter_GPT2LMHeadModel_WithoutOptionalModel_HF_CS21(
BaseCheckpointConverter_HF_CS
):
def __init__(self):
super().__init__()
self.rules = [
ConversionRule(
["lm_head\.(?:weight|bias)"],
action=self.replaceKey,
),
ConversionRule(
[
EquivalentSubkey("transformer.", ""),
Converter_GPT2Model_WithoutOptionalModel_HF_CS21(),
],
action=None,
),
]
@staticmethod
def formats() -> Tuple[FormatVersions, FormatVersions]:
return (
FormatVersions("hf"),
FormatVersions("cs-2.1", "cs-2.2", "cs-2.3", "cs-2.4"),
)
@staticmethod
def get_config_converter_class() -> BaseConfigConverter:
return ConfigConverter_GPT2Model_HF_CS21
def supports_mup_conversion(self) -> bool:
return True
Converter_GPT2LMHeadModel_HF_CS21 = Build_HF_CS_Converter_WithOptionalModel(
"Converter_GPT2LMHeadModel_HF_CS21",
Converter_GPT2LMHeadModel_WithoutOptionalModel_HF_CS21,
derived_class=Converter_GPT2LMHeadModel_WithoutOptionalModel_HF_CS21,
converter_note_fn=lambda cls: "{} GPT2LMHeadModel <-> {} GPT2LMHeadModel".format(
cls.formats()[0], cls.formats()[1]
),
)
[docs]class Converter_GPT2LMHeadModel_CS22_CS23(BaseCheckpointConverter_CS_CS):
def __init__(self):
super().__init__()
self.rules = [
ConversionRule([".*"], action=self.replaceKey),
]
@classmethod
def converter_note(cls) -> str:
return "GPT2LMHeadModel class"
@staticmethod
def formats() -> Tuple[FormatVersions, FormatVersions]:
return (FormatVersions("cs-2.2"), FormatVersions("cs-2.3", "cs-2.4"))
@staticmethod
def get_config_converter_class() -> BaseConfigConverter:
return ConfigConverter_GPT2Model_CS22_CS23
[docs]class ConfigConverter_GPT2Model_CS22_CS23(BaseConfigConverter_CS_CS):
def __init__(self):
super().__init__()
self.rules = [
# The following params were deprecated:
ConversionRule(["use_position_embedding"], action=None),
ConversionRule(["alibi_implementation"], action=None),
ConversionRule(["weight_initialization_seed"], action=None),
# Remaining params can be copied:
ConversionRule([".*"], action=self.replaceKey),
]
@staticmethod
def formats() -> Tuple[FormatVersions, FormatVersions]:
return (FormatVersions("cs-2.2"), FormatVersions("cs-2.3", "cs-2.4"))