convert_mistral_moe_weights_to_hf.py

# Copyright 2023 Mistral AI 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 argparse
import gc
import json
import os
import shutil
import warnings

import torch

from transformers import (
    LlamaTokenizer
)

from transformers.models.mistral.modeling_moe_mistral import MixtralForCausalLM
from transformers.models.mistral.configuration_dmoe_mistral import MixtralConfig

try:
    from transformers import LlamaTokenizerFast

    tokenizer_class = LlamaTokenizerFast
except ImportError as e:
    warnings.warn(e)
    warnings.warn(
        "The converted tokenizer will be the `slow` tokenizer. To use the fast, update your `tokenizers` library and re-run the tokenizer conversion"
    )
    tokenizer_class = LlamaTokenizer

"""
Sample usage:

```
python src/transformers/models/mistral/convert_mistral_weights_to_hf.py \
    --input_dir /path/to/downloaded/mistral/weights --model_size 7B --output_dir /output/path
```

Thereafter, models can be loaded via:

```py
from transformers import MistralForCausalLM, LlamaTokenizer

model = MistralForCausalLM.from_pretrained("/output/path")
tokenizer = LlamaTokenizer.from_pretrained("/output/path")
```

Important note: you need to be able to host the whole model in RAM to execute this script (even if the biggest versions
come in several checkpoints they each contain a part of each weight of the model, so we need to load them all in RAM).
"""

NUM_SHARDS = {"7B": 1}


def compute_intermediate_size(n, ffn_dim_multiplier=1, multiple_of=256):
    return multiple_of * ((int(ffn_dim_multiplier * int(8 * n / 3)) + multiple_of - 1) // multiple_of)


def read_json(path):
    with open(path, "r") as f:
        return json.load(f)


def write_json(text, path):
    with open(path, "w") as f:
        json.dump(text, f)


def write_model(model_path, input_base_path, model_size, tokenizer_path=None, safe_serialization=True):
    # for backward compatibility, before you needed the repo to be called `my_repo/model_size`
    if not os.path.isfile(os.path.join(input_base_path, "params.json")):
        input_base_path = os.path.join(input_base_path, model_size)

    os.makedirs(model_path, exist_ok=True)
    tmp_model_path = os.path.join(model_path, "tmp")
    os.makedirs(tmp_model_path, exist_ok=True)

    params = read_json(os.path.join(input_base_path, "params.json"))
    num_shards = NUM_SHARDS[model_size]

    n_layers = params["n_layers"]
    n_heads = params["n_heads"]
    n_heads_per_shard = n_heads // num_shards
    dim = params["dim"]
    dims_per_head = dim // n_heads
    base = params.get("rope_theta", 100000.0)
    inv_freq = 1.0 / (base ** (torch.arange(0, dims_per_head, 2).float() / dims_per_head))
    max_position_embeddings = 4096 * 8
    num_experts_per_token = params["moe"]["num_experts_per_tok"]
    num_experts = params["moe"]["num_experts"]
    

    if tokenizer_path is not None:
        tokenizer = tokenizer_class(tokenizer_path)
        tokenizer.save_pretrained(model_path)
    vocab_size = tokenizer.vocab_size if tokenizer_path is not None else 32000

    if "n_kv_heads" in params:
        num_key_value_heads = params["n_kv_heads"]  # for GQA / MQA
        num_local_key_value_heads = num_key_value_heads // num_shards
        key_value_dim = dims_per_head * num_local_key_value_heads
    else:  # compatibility with other checkpoints
        num_key_value_heads = n_heads
        num_local_key_value_heads = n_heads_per_shard
        key_value_dim = dim

    # permute for sliced rotary
    def permute(w, n_heads=n_heads, dim1=dim, dim2=dim):
        return w.view(n_heads, dim1 // n_heads // 2, 2, dim2).transpose(1, 2).reshape(dim1, dim2)
    

    print(f"Fetching all parameters from the checkpoint at {input_base_path}.")
    # Load weights
    loaded = [
        torch.load(os.path.join(input_base_path, f"consolidated.{i:02d}.pth"), map_location="cpu")
        for i in range(num_shards)
    ]
    param_count = 0
    index_dict = {"weight_map": {}}
    for layer_i in range(n_layers):
        filename = f"pytorch_model-{layer_i + 1}-of-{n_layers + 1}.bin"

        # Sharded
        # Note that attention.w{q,k,v,o}, feed_fordward.w[1,2,3], attention_norm.weight and ffn_norm.weight share
        # the same storage object, saving attention_norm and ffn_norm will save other weights too, which is
        # redundant as other weights will be stitched from multiple shards. To avoid that, they are cloned.

        state_dict = {
            f"model.layers.{layer_i}.input_layernorm.weight": loaded[0][
                f"layers.{layer_i}.attention_norm.weight"
            ].clone(),
            f"model.layers.{layer_i}.post_attention_layernorm.weight": loaded[0][
                f"layers.{layer_i}.ffn_norm.weight"
            ].clone(),
        }
        state_dict[f"model.layers.{layer_i}.self_attn.q_proj.weight"] = permute(
            torch.cat(
                [
                    loaded[i][f"layers.{layer_i}.attention.wq.weight"].view(n_heads_per_shard, dims_per_head, dim)
                    for i in range(num_shards)
                ],
                dim=0,
            ).reshape(dim, dim)
        )
        state_dict[f"model.layers.{layer_i}.self_attn.k_proj.weight"] = permute(
            torch.cat(
                [
                    loaded[i][f"layers.{layer_i}.attention.wk.weight"].view(
                        num_local_key_value_heads, dims_per_head, dim
                    )
                    for i in range(num_shards)
                ],
                dim=0,
            ).reshape(key_value_dim, dim),
            num_key_value_heads,
            key_value_dim,
            dim,
        )
        state_dict[f"model.layers.{layer_i}.self_attn.v_proj.weight"] = torch.cat(
            [
                loaded[i][f"layers.{layer_i}.attention.wv.weight"].view(num_local_key_value_heads, dims_per_head, dim)
                for i in range(num_shards)
            ],
            dim=0,
        ).reshape(key_value_dim, dim)

        state_dict[f"model.layers.{layer_i}.self_attn.o_proj.weight"] = torch.cat(
            [loaded[i][f"layers.{layer_i}.attention.wo.weight"] for i in range(num_shards)], dim=1
        )

        for expert in range(num_experts):
            state_dict[f"model.layers.{layer_i}.mlp.experts.{expert}.w1.weight"] = loaded[0][f"layers.{layer_i}.feed_forward.experts.{expert}.w1.weight"]
            state_dict[f"model.layers.{layer_i}.mlp.experts.{expert}.w2.weight"] = loaded[0][f"layers.{layer_i}.feed_forward.experts.{expert}.w2.weight"]
            state_dict[f"model.layers.{layer_i}.mlp.experts.{expert}.w3.weight"] = loaded[0][f"layers.{layer_i}.feed_forward.experts.{expert}.w3.weight"] 

        state_dict[f"model.layers.{layer_i}.mlp.gate.weight"] = loaded[0][f"layers.{layer_i}.feed_forward.gate.weight"]

        state_dict[f"model.layers.{layer_i}.self_attn.rotary_emb.inv_freq"] = inv_freq
        for k, v in state_dict.items():
            index_dict["weight_map"][k] = filename
            param_count += v.numel()
        torch.save(state_dict, os.path.join(tmp_model_path, filename))

    filename = f"pytorch_model-{n_layers + 1}-of-{n_layers + 1}.bin"
    state_dict = {
        "model.norm.weight": loaded[0]["norm.weight"],
        "model.embed_tokens.weight": torch.cat([loaded[i]["tok_embeddings.weight"] for i in range(num_shards)], dim=1),
        "lm_head.weight": torch.cat([loaded[i]["output.weight"] for i in range(num_shards)], dim=0),
    }

    for k, v in state_dict.items():
        index_dict["weight_map"][k] = filename
        param_count += v.numel()
        print(param_count)
    torch.save(state_dict, os.path.join(tmp_model_path, filename))

    index_dict["metadata"] = {"total_size": param_count * 2}
    write_json(index_dict, os.path.join(tmp_model_path, "pytorch_model.bin.index.json"))
    config = MixtralConfig(
        hidden_size=dim,
        intermediate_size=params["hidden_dim"],
        num_attention_heads=params["n_heads"],
        num_hidden_layers=params["n_layers"],
        rms_norm_eps=params["norm_eps"],
        num_key_value_heads=num_key_value_heads,
        vocab_size=vocab_size,
        rope_theta=base,
        max_position_embeddings=max_position_embeddings,
        num_experts=num_experts,
        num_experts_per_token=num_experts_per_token
    )
    config.save_pretrained(tmp_model_path)

    del state_dict
    del loaded
    gc.collect()

    print("Loading the checkpoint in a Mistral model.")
    model = MixtralForCausalLM.from_pretrained(tmp_model_path, torch_dtype=torch.float16, low_cpu_mem_usage=True)
    # Avoid saving this as part of the config.
    del model.config._name_or_path
    model.config.torch_dtype = torch.float16
    print("Saving in the Transformers format.")
    model.save_pretrained(model_path, safe_serialization=safe_serialization)
    shutil.rmtree(tmp_model_path)


def write_tokenizer(tokenizer_path, input_tokenizer_path):
    # Initialize the tokenizer based on the `spm` model
    print(f"Saving a {tokenizer_class.__name__} to {tokenizer_path}.")
    tokenizer = tokenizer_class(input_tokenizer_path)
    tokenizer.save_pretrained(tokenizer_path)


def main():
    parser = argparse.ArgumentParser()
    parser.add_argument(
        "--input_dir",
        help="Location of Mistral weights, which contains tokenizer.model and model folders",
    )
    parser.add_argument(
        "--model_size",
        choices=["7B", "tokenizer_only"],
        help="'f' models correspond to the finetuned versions, and are specific to the Mistral2 official release. For more details on Mistral2, checkout the original repo: https://huggingface.co/meta-mistral",
    )
    parser.add_argument(
        "--output_dir",
        help="Location to write HF model and tokenizer",
    )
    parser.add_argument("--safe_serialization", type=bool, help="Whether or not to save using `safetensors`.")
    args = parser.parse_args()
    spm_path = os.path.join(args.input_dir, "tokenizer.model")
    if args.model_size != "tokenizer_only":
        write_model(
            model_path=args.output_dir,
            input_base_path=args.input_dir,
            model_size=args.model_size,
            safe_serialization=args.safe_serialization,
            tokenizer_path=spm_path,
        )
    else:
        write_tokenizer(args.output_dir, spm_path)


if __name__ == "__main__":
    main()