fixed imports models.llama3. --> models.llama3_1.api.

toolchain/common/deployment_types.py

@@ -1,7 +1,7 @@
 from enum import Enum
 from typing import Dict, Optional
 
-from models.llama3.datatypes import URL
+from models.llama3_1.api.datatypes import URL
 
 from pydantic import BaseModel
 

toolchain/common/training_types.py

@@ -1,4 +1,4 @@
-from models.llama3.datatypes import URL
+from models.llama3_1.api.datatypes import URL
 from pydantic import BaseModel 
 
 

toolchain/dataset/api/datatypes.py

@@ -4,7 +4,7 @@ from typing import Any, Dict, Optional
 from pydantic import BaseModel
 
 from strong_typing.schema import json_schema_type
-from models.llama3.datatypes import URL
+from models.llama3_1.api.datatypes import URL
 
 
 @json_schema_type

toolchain/evaluations/api/endpoints.py

@@ -4,7 +4,7 @@ from pydantic import BaseModel
 
 from pyopenapi import webmethod
 
-from models.llama3.datatypes import * # noqa: F403
+from models.llama3_1.api.datatypes import * # noqa: F403
 from .datatypes import *  # noqa: F403
 from toolchain.dataset.api.datatypes import *  # noqa: F403
 from toolchain.common.training_types import *  # noqa: F403

toolchain/inference/api/datatypes.py

@@ -6,7 +6,7 @@ from pydantic import BaseModel, Field
 from strong_typing.schema import json_schema_type
 from typing_extensions import Annotated
 
-from models.llama3.datatypes import *  # noqa: F403
+from models.llama3_1.api.datatypes import *  # noqa: F403
 
 
 class LogProbConfig(BaseModel):

toolchain/inference/generation.py

@@ -16,11 +16,11 @@ from fairscale.nn.model_parallel.initialize import (
     initialize_model_parallel,
     model_parallel_is_initialized,
 )
-from models.llama3.args import ModelArgs
+from models.llama3_1.api.args import ModelArgs
-from models.llama3.chat_format import ChatFormat, ModelInput
+from models.llama3_1.api.chat_format import ChatFormat, ModelInput
-from models.llama3.datatypes import Message
+from models.llama3_1.api.datatypes import Message
-from models.llama3.model import Transformer
+from models.llama3_1.api.model import Transformer
-from models.llama3.tokenizer import Tokenizer
+from models.llama3_1.api.tokenizer import Tokenizer
 from termcolor import cprint
 
 

toolchain/inference/inference.py

@@ -1,6 +1,6 @@
 from typing import AsyncGenerator
 
-from models.llama3.datatypes import StopReason
+from models.llama3_1.api.datatypes import StopReason
 
 from .api.config import CheckpointType, GeneratorArgs, InlineImplConfig
 from .api.datatypes import (

toolchain/inference/model_parallel.py

@@ -2,9 +2,9 @@ from dataclasses import dataclass
 from functools import partial
 from typing import Generator, List, Optional
 
-from models.llama3.chat_format import ChatFormat
+from models.llama3_1.api.chat_format import ChatFormat
-from models.llama3.datatypes import Message
+from models.llama3_1.api.datatypes import Message
-from models.llama3.tokenizer import Tokenizer
+from models.llama3_1.api.tokenizer import Tokenizer
 
 from .api.config import GeneratorArgs
 from .generation import Llama

toolchain/post_training/api/endpoints.py

@@ -7,7 +7,7 @@ from pydantic import BaseModel, Field
 from pyopenapi import webmethod
 from strong_typing.schema import json_schema_type
 
-from models.llama3.datatypes import *  # noqa: F403
+from models.llama3_1.api.datatypes import *  # noqa: F403
 from toolchain.dataset.api.datatypes import *  # noqa: F403
 from toolchain.common.training_types import *  # noqa: F403
 from .datatypes import *  # noqa: F403

toolchain/reward_scoring/api/datatypes.py

@@ -4,7 +4,7 @@ from pydantic import BaseModel
 
 from strong_typing.schema import json_schema_type
 
-from models.llama3.datatypes import *  # noqa: F403
+from models.llama3_1.api.datatypes import *  # noqa: F403
 
 
 @json_schema_type

toolchain/safety/api/datatypes.py

@@ -1,7 +1,7 @@
 from enum import Enum
 from typing import Dict, Optional, Union
 
-from models.llama3.datatypes import ToolParamDefinition
+from models.llama3_1.api.datatypes import ToolParamDefinition
 
 from pydantic import BaseModel
 

toolchain/safety/shields/base.py

@@ -1,7 +1,7 @@
 from abc import ABC, abstractmethod
 from typing import List, Union
 
-from models.llama3.datatypes import Attachment, Message
+from models.llama3_1.api.datatypes import Attachment, Message
 from toolchain.safety.api.datatypes import *  # noqa: F403
 
 CANNED_RESPONSE_TEXT = "I can't answer that. Can I help with something else?"

toolchain/safety/shields/contrib/third_party_shield.py

@@ -1,7 +1,7 @@
 import sys
 from typing import List
 
-from models.llama3.datatypes import Message
+from models.llama3_1.api.datatypes import Message
 
 parent_dir = "../.."
 sys.path.append(parent_dir)

toolchain/safety/shields/llama_guard.py

@@ -4,7 +4,7 @@ from string import Template
 from typing import List, Optional
 
 import torch
-from models.llama3.datatypes import Message
+from models.llama3_1.api.datatypes import Message
 from termcolor import cprint
 from transformers import AutoModelForCausalLM, AutoTokenizer
 

toolchain/safety/shields/prompt_guard.py

@@ -3,7 +3,7 @@ from typing import List
 
 import torch
 
-from models.llama3.datatypes import Message
+from models.llama3_1.api.datatypes import Message
 from termcolor import cprint
 from transformers import AutoModelForSequenceClassification, AutoTokenizer
 

toolchain/safety/shields/shield_runner.py

@@ -1,7 +1,7 @@
 import asyncio
 from typing import List
 
-from models.llama3.datatypes import Message, Role
+from models.llama3_1.api.datatypes import Message, Role
 
 from .base import OnViolationAction, ShieldBase, ShieldResponse
 

toolchain/spec/generate.py

@@ -4,7 +4,7 @@ import yaml
 
 from pyopenapi import Info, Options, Server, Specification
 
-from models.llama3.datatypes import *  # noqa: F403
+from models.llama3_1.api.datatypes import *  # noqa: F403
 from toolchain.dataset.api import *  # noqa: F403
 from toolchain.evaluations.api import *  # noqa: F403
 from toolchain.inference.api import *  # noqa: F403

toolchain/spec/package.sh

@@ -7,7 +7,7 @@ echo "Using temporary directory: $TMPDIR"
 
 rootdir=$(git rev-parse --show-toplevel)
 
-files_to_copy=("toolchain/spec/openapi*" "models/llama3/datatypes.py" "toolchain/inference/api/*.py" "agentic_system/api/*.py" "toolchain/common/*.py" "toolchain/dataset/api/*.py" "toolchain/evaluations/api/*.py" "toolchain/reward_scoring/api/*.py" "toolchain/post_training/api/*.py" "toolchain/safety/api/*.py")
+files_to_copy=("toolchain/spec/openapi*" "models.llama3_1.api.datatypes.py" "toolchain/inference/api/*.py" "agentic_system/api/*.py" "toolchain/common/*.py" "toolchain/dataset/api/*.py" "toolchain/evaluations/api/*.py" "toolchain/reward_scoring/api/*.py" "toolchain/post_training/api/*.py" "toolchain/safety/api/*.py")
 for file in "${files_to_copy[@]}"; do
     relpath="$file"
     set -x 

toolchain/spec/post_training_types.py

@@ -1,7 +1,7 @@
 from enum import Enum
 from typing import Any, Dict, List
 
-from models.llama3.datatypes import URL
+from models.llama3_1.api.datatypes import URL
 from pydantic import BaseModel, Field
 
 from strong_typing.schema import json_schema_type

toolchain/synthetic_data_generation/api/endpoints.py

@@ -5,7 +5,7 @@ from pydantic import BaseModel
 from pyopenapi import webmethod
 from strong_typing.schema import json_schema_type
 
-from models.llama3.datatypes import *  # noqa: F403
+from models.llama3_1.api.datatypes import *  # noqa: F403
 from toolchain.reward_scoring.api.datatypes import *  # noqa: F403
 from .datatypes import *  # noqa: F403