llama.cpp/llama_cpp/llama.py

2009 lines
79 KiB
Python
Raw Normal View History

2024-01-17 14:16:13 +00:00
from __future__ import annotations
2023-03-24 19:47:17 +00:00
import os
2023-04-04 17:09:24 +00:00
import sys
2023-03-23 09:33:06 +00:00
import uuid
import time
import json
import ctypes
import fnmatch
2023-03-23 09:33:06 +00:00
import multiprocessing
from typing import (
List,
Optional,
Union,
Generator,
Sequence,
Iterator,
Deque,
Callable,
Dict,
)
from collections import deque
from pathlib import Path
2023-03-23 09:33:06 +00:00
from llama_cpp.llama_types import List
2023-04-01 17:01:27 +00:00
from .llama_types import *
2023-08-06 17:21:37 +00:00
from .llama_grammar import LlamaGrammar
from .llama_cache import (
BaseLlamaCache,
LlamaCache, # type: ignore
LlamaDiskCache, # type: ignore
LlamaRAMCache, # type: ignore
)
from .llama_tokenizer import BaseLlamaTokenizer, LlamaTokenizer
import llama_cpp.llama_cpp as llama_cpp
import llama_cpp.llama_chat_format as llama_chat_format
2023-03-23 09:33:06 +00:00
from llama_cpp.llama_speculative import LlamaDraftModel
2023-05-26 20:12:45 +00:00
import numpy as np
import numpy.typing as npt
from ._internals import (
_LlamaModel, # type: ignore
_LlamaContext, # type: ignore
_LlamaBatch, # type: ignore
_LlamaTokenDataArray, # type: ignore
_LlamaSamplingParams, # type: ignore
_LlamaSamplingContext, # type: ignore
)
from ._logger import set_verbose
from ._utils import suppress_stdout_stderr
2023-09-29 02:42:03 +00:00
2023-03-23 09:33:06 +00:00
class Llama:
2023-03-24 22:57:59 +00:00
"""High-level Python wrapper for a llama.cpp model."""
__backend_initialized = False
2023-03-23 09:33:06 +00:00
def __init__(
self,
model_path: str,
*,
2023-09-29 02:42:03 +00:00
# Model Params
n_gpu_layers: int = 0,
2024-02-25 21:53:58 +00:00
split_mode: int = llama_cpp.LLAMA_SPLIT_MODE_LAYER,
main_gpu: int = 0,
tensor_split: Optional[List[float]] = None,
2023-09-29 02:42:03 +00:00
vocab_only: bool = False,
use_mmap: bool = True,
use_mlock: bool = False,
kv_overrides: Optional[Dict[str, Union[bool, int, float]]] = None,
2023-09-29 02:42:03 +00:00
# Context Params
seed: int = llama_cpp.LLAMA_DEFAULT_SEED,
n_ctx: int = 512,
n_batch: int = 512,
n_threads: Optional[int] = None,
n_threads_batch: Optional[int] = None,
2024-02-25 21:53:58 +00:00
rope_scaling_type: Optional[int] = llama_cpp.LLAMA_ROPE_SCALING_TYPE_UNSPECIFIED,
pooling_type: int = llama_cpp.LLAMA_POOLING_TYPE_UNSPECIFIED,
rope_freq_base: float = 0.0,
rope_freq_scale: float = 0.0,
2023-11-03 15:34:50 +00:00
yarn_ext_factor: float = -1.0,
yarn_attn_factor: float = 1.0,
yarn_beta_fast: float = 32.0,
yarn_beta_slow: float = 1.0,
yarn_orig_ctx: int = 0,
2023-03-23 09:33:06 +00:00
logits_all: bool = False,
2023-03-25 20:26:23 +00:00
embedding: bool = False,
2024-01-18 16:08:57 +00:00
offload_kqv: bool = True,
2023-09-29 02:42:03 +00:00
# Sampling Params
2023-04-01 17:01:27 +00:00
last_n_tokens_size: int = 64,
2023-09-29 02:42:03 +00:00
# LoRA Params
lora_base: Optional[str] = None,
2023-09-29 02:42:03 +00:00
lora_scale: float = 1.0,
lora_path: Optional[str] = None,
2023-09-29 02:42:03 +00:00
# Backend Params
2024-02-17 05:37:51 +00:00
numa: Union[bool, int] = False,
# Chat Format Params
chat_format: Optional[str] = None,
chat_handler: Optional[llama_chat_format.LlamaChatCompletionHandler] = None,
# Speculative Decoding
draft_model: Optional[LlamaDraftModel] = None,
# Tokenizer Override
tokenizer: Optional[BaseLlamaTokenizer] = None,
# KV cache quantization
type_k: Optional[int] = None,
type_v: Optional[int] = None,
2023-09-29 02:42:03 +00:00
# Misc
2023-04-04 17:09:24 +00:00
verbose: bool = True,
2023-09-29 02:42:03 +00:00
# Extra Params
**kwargs, # type: ignore
2023-04-01 17:01:27 +00:00
):
2023-03-24 22:57:59 +00:00
"""Load a llama.cpp model from `model_path`.
2023-11-23 04:10:04 +00:00
Examples:
Basic usage
>>> import llama_cpp
>>> model = llama_cpp.Llama(
... model_path="path/to/model",
... )
>>> print(model("The quick brown fox jumps ", stop=["."])["choices"][0]["text"])
the lazy dog
Loading a chat model
>>> import llama_cpp
>>> model = llama_cpp.Llama(
... model_path="path/to/model",
... chat_format="llama-2",
... )
>>> print(model.create_chat_completion(
... messages=[{
... "role": "user",
... "content": "what is the meaning of life?"
... }]
... ))
2023-03-24 22:57:59 +00:00
Args:
2023-03-25 16:33:18 +00:00
model_path: Path to the model.
n_gpu_layers: Number of layers to offload to GPU (-ngl). If -1, all layers are offloaded.
2024-01-15 17:49:20 +00:00
split_mode: How to split the model across GPUs. See llama_cpp.LLAMA_SPLIT_* for options.
main_gpu: main_gpu interpretation depends on split_mode: LLAMA_SPLIT_NONE: the GPU that is used for the entire model. LLAMA_SPLIT_ROW: the GPU that is used for small tensors and intermediate results. LLAMA_SPLIT_LAYER: ignored
tensor_split: How split tensors should be distributed across GPUs. If None, the model is not split.
vocab_only: Only load the vocabulary no weights.
use_mmap: Use mmap if possible.
use_mlock: Force the system to keep the model in RAM.
kv_overrides: Key-value overrides for the model.
2023-11-26 20:56:40 +00:00
seed: RNG seed, -1 for random
n_ctx: Text context, 0 = from model
n_batch: Prompt processing maximum batch size
n_threads: Number of threads to use for generation
n_threads_batch: Number of threads to use for batch processing
rope_scaling_type: RoPE scaling type, from `enum llama_rope_scaling_type`. ref: https://github.com/ggerganov/llama.cpp/pull/2054
pooling_type: Pooling type, from `enum llama_pooling_type`.
2023-11-26 20:56:40 +00:00
rope_freq_base: RoPE base frequency, 0 = from model
rope_freq_scale: RoPE frequency scaling factor, 0 = from model
yarn_ext_factor: YaRN extrapolation mix factor, negative = from model
yarn_attn_factor: YaRN magnitude scaling factor
yarn_beta_fast: YaRN low correction dim
yarn_beta_slow: YaRN high correction dim
yarn_orig_ctx: YaRN original context size
logits_all: Return logits for all tokens, not just the last token. Must be True for completion to return logprobs.
2023-03-25 20:26:23 +00:00
embedding: Embedding mode only.
offload_kqv: Offload K, Q, V to GPU.
2023-04-01 17:01:27 +00:00
last_n_tokens_size: Maximum number of tokens to keep in the last_n_tokens deque.
lora_base: Optional path to base model, useful if using a quantized base model and you want to apply LoRA to an f16 model.
2023-04-18 05:43:44 +00:00
lora_path: Path to a LoRA file to apply to the model.
2024-02-17 05:37:51 +00:00
numa: numa policy
chat_format: String specifying the chat format to use when calling create_chat_completion.
chat_handler: Optional chat handler to use when calling create_chat_completion.
draft_model: Optional draft model to use for speculative decoding.
tokenizer: Optional tokenizer to override the default tokenizer from llama.cpp.
2023-04-04 17:09:24 +00:00
verbose: Print verbose output to stderr.
type_k: KV cache data type for K (default: f16)
type_v: KV cache data type for V (default: f16)
2023-03-24 22:57:59 +00:00
Raises:
ValueError: If the model path does not exist.
Returns:
A Llama instance.
"""
2023-04-04 17:09:24 +00:00
self.verbose = verbose
set_verbose(verbose)
if not Llama.__backend_initialized:
with suppress_stdout_stderr(disable=verbose):
2024-02-17 05:37:51 +00:00
llama_cpp.llama_backend_init()
Llama.__backend_initialized = True
2024-02-17 05:37:51 +00:00
if isinstance(numa, bool):
self.numa = (
llama_cpp.GGML_NUMA_STRATEGY_DISTRIBUTE
if numa
else llama_cpp.GGML_NUMA_STRATEGY_DISABLED
)
2024-02-17 06:02:33 +00:00
else:
self.numa = numa
2024-02-17 05:37:51 +00:00
if self.numa != llama_cpp.GGML_NUMA_STRATEGY_DISABLED:
with suppress_stdout_stderr(disable=verbose):
llama_cpp.llama_numa_init(self.numa)
2023-03-23 09:33:06 +00:00
self.model_path = model_path
2023-09-29 02:42:03 +00:00
# Model Params
self.model_params = llama_cpp.llama_model_default_params()
self.model_params.n_gpu_layers = (
0x7FFFFFFF if n_gpu_layers == -1 else n_gpu_layers
) # 0x7FFFFFFF is INT32 max, will be auto set to all layers
2024-01-15 17:49:20 +00:00
self.model_params.split_mode = split_mode
2023-09-29 02:42:03 +00:00
self.model_params.main_gpu = main_gpu
2023-07-15 19:11:01 +00:00
self.tensor_split = tensor_split
2023-12-22 20:12:27 +00:00
self._c_tensor_split = None
2023-07-15 19:11:01 +00:00
if self.tensor_split is not None:
if len(self.tensor_split) > llama_cpp.LLAMA_MAX_DEVICES:
2023-11-06 14:16:36 +00:00
raise ValueError(
f"Attempt to split tensors that exceed maximum supported devices. Current LLAMA_MAX_DEVICES={llama_cpp.LLAMA_MAX_DEVICES}"
)
# Type conversion and expand the list to the length of LLAMA_MAX_DEVICES
2023-09-14 00:00:42 +00:00
FloatArray = ctypes.c_float * llama_cpp.LLAMA_MAX_DEVICES
self._c_tensor_split = FloatArray(
2023-09-29 02:42:03 +00:00
*tensor_split # type: ignore
) # keep a reference to the array so it is not gc'd
2023-09-29 02:42:03 +00:00
self.model_params.tensor_split = self._c_tensor_split
self.model_params.vocab_only = vocab_only
self.model_params.use_mmap = use_mmap if lora_path is None else False
self.model_params.use_mlock = use_mlock
2023-07-15 19:11:01 +00:00
# kv_overrides is the original python dict
self.kv_overrides = kv_overrides
if kv_overrides is not None:
# _kv_overrides_array is a ctypes.Array of llama_model_kv_override Structs
2024-01-24 03:08:27 +00:00
kvo_array_len = len(kv_overrides) + 1 # for sentinel element
self._kv_overrides_array = (
llama_cpp.llama_model_kv_override * kvo_array_len
)()
for i, (k, v) in enumerate(kv_overrides.items()):
2024-01-24 03:08:27 +00:00
self._kv_overrides_array[i].key = k.encode("utf-8")
2024-01-24 03:28:03 +00:00
if isinstance(v, bool):
2024-02-25 21:53:58 +00:00
self._kv_overrides_array[i].tag = llama_cpp.LLAMA_KV_OVERRIDE_TYPE_BOOL
2024-01-24 03:28:03 +00:00
self._kv_overrides_array[i].value.bool_value = v
elif isinstance(v, int):
2024-02-25 21:53:58 +00:00
self._kv_overrides_array[i].tag = llama_cpp.LLAMA_KV_OVERRIDE_TYPE_INT
self._kv_overrides_array[i].value.int_value = v
elif isinstance(v, float):
2024-02-25 21:53:58 +00:00
self._kv_overrides_array[i].tag = llama_cpp.LLAMA_KV_OVERRIDE_TYPE_FLOAT
self._kv_overrides_array[i].value.float_value = v
else:
raise ValueError(f"Unknown value type for {k}: {v}")
self._kv_overrides_array[-1].key = (
b"\0" # ensure sentinel element is zeroed
)
self.model_params.kv_overrides = self._kv_overrides_array
2023-09-29 02:42:03 +00:00
self.n_batch = min(n_ctx, n_batch) # ???
self.n_threads = n_threads or max(multiprocessing.cpu_count() // 2, 1)
self.n_threads_batch = n_threads_batch or multiprocessing.cpu_count()
2023-09-29 02:42:03 +00:00
# Context Params
self.context_params = llama_cpp.llama_context_default_params()
self.context_params.seed = seed
self.context_params.n_ctx = n_ctx
self.context_params.n_batch = self.n_batch
self.context_params.n_threads = self.n_threads
self.context_params.n_threads_batch = self.n_threads_batch
self.context_params.rope_scaling_type = (
2023-11-06 14:16:36 +00:00
rope_scaling_type
if rope_scaling_type is not None
2024-02-25 21:53:58 +00:00
else llama_cpp.LLAMA_ROPE_SCALING_TYPE_UNSPECIFIED
)
self.context_params.pooling_type = pooling_type
self.context_params.rope_freq_base = (
rope_freq_base if rope_freq_base != 0.0 else 0
)
self.context_params.rope_freq_scale = (
rope_freq_scale if rope_freq_scale != 0.0 else 0
)
self.context_params.yarn_ext_factor = (
yarn_ext_factor if yarn_ext_factor != 0.0 else 0
)
self.context_params.yarn_attn_factor = (
yarn_attn_factor if yarn_attn_factor != 0.0 else 0
)
self.context_params.yarn_beta_fast = (
yarn_beta_fast if yarn_beta_fast != 0.0 else 0
)
self.context_params.yarn_beta_slow = (
yarn_beta_slow if yarn_beta_slow != 0.0 else 0
)
2023-11-06 14:16:36 +00:00
self.context_params.yarn_orig_ctx = yarn_orig_ctx if yarn_orig_ctx != 0 else 0
self.context_params.logits_all = (
logits_all if draft_model is None else True
) # Must be set to True for speculative decoding
2024-03-06 06:32:00 +00:00
self.context_params.embeddings = embedding # TODO: Rename to embeddings
self.context_params.offload_kqv = offload_kqv
# KV cache quantization
if type_k is not None:
self.context_params.type_k = type_k
if type_v is not None:
self.context_params.type_v = type_v
2023-09-29 02:42:03 +00:00
# Sampling Params
2023-04-01 17:01:27 +00:00
self.last_n_tokens_size = last_n_tokens_size
2023-03-23 09:33:06 +00:00
2023-09-29 02:42:03 +00:00
self.cache: Optional[BaseLlamaCache] = None
2023-03-23 09:33:06 +00:00
self.lora_base = lora_base
2023-09-29 02:42:03 +00:00
self.lora_scale = lora_scale
self.lora_path = lora_path
2023-03-24 19:47:17 +00:00
if not os.path.exists(model_path):
raise ValueError(f"Model path does not exist: {model_path}")
2023-11-06 14:16:36 +00:00
self._model = _LlamaModel(
path_model=self.model_path, params=self.model_params, verbose=self.verbose
)
# Override tokenizer
self.tokenizer_ = tokenizer or LlamaTokenizer(self)
# Set the default value for the context and correct the batch
if n_ctx == 0:
n_ctx = self._model.n_ctx_train()
self.n_batch = min(n_ctx, n_batch)
self.context_params.n_ctx = self._model.n_ctx_train()
self.context_params.n_batch = self.n_batch
2023-03-23 09:33:06 +00:00
2023-11-06 14:16:36 +00:00
self._ctx = _LlamaContext(
model=self._model,
params=self.context_params,
verbose=self.verbose,
)
2023-11-06 14:16:36 +00:00
self._batch = _LlamaBatch(
n_tokens=self.n_batch,
embd=0,
n_seq_max=self.context_params.n_ctx,
verbose=self.verbose,
)
if self.lora_path:
2023-11-06 14:16:36 +00:00
if self._model.apply_lora_from_file(
self.lora_path,
2023-09-29 02:42:03 +00:00
self.lora_scale,
2023-11-06 14:16:36 +00:00
self.lora_base,
self.n_threads,
2023-04-18 05:43:44 +00:00
):
raise RuntimeError(
f"Failed to apply LoRA from lora path: {self.lora_path} to base path: {self.lora_base}"
)
2023-03-23 09:33:06 +00:00
2023-04-04 17:09:24 +00:00
if self.verbose:
print(llama_cpp.llama_print_system_info().decode("utf-8"), file=sys.stderr)
self.chat_format = chat_format
self.chat_handler = chat_handler
2023-04-04 17:09:24 +00:00
self.draft_model = draft_model
2023-05-23 21:56:21 +00:00
self._n_vocab = self.n_vocab()
self._n_ctx = self.n_ctx()
2023-11-06 14:16:36 +00:00
2023-08-24 04:17:00 +00:00
self._token_nl = self.token_nl()
self._token_eos = self.token_eos()
2023-11-06 14:16:36 +00:00
self._candidates = _LlamaTokenDataArray(n_vocab=self._n_vocab)
2023-04-04 17:09:24 +00:00
self.n_tokens = 0
self.input_ids: npt.NDArray[np.intc] = np.ndarray((n_ctx,), dtype=np.intc)
self.scores: npt.NDArray[np.single] = np.ndarray(
(n_ctx, self._n_vocab), dtype=np.single
)
2024-01-24 03:08:27 +00:00
self._mirostat_mu = ctypes.c_float(
2.0 * 5.0
) # TODO: Move this to sampling context
2024-01-19 13:31:59 +00:00
try:
self.metadata = self._model.metadata()
except Exception as e:
self.metadata = {}
if self.verbose:
print(f"Failed to load metadata: {e}", file=sys.stderr)
2024-01-24 03:08:27 +00:00
if self.verbose:
print(f"Model metadata: {self.metadata}", file=sys.stderr)
if (
self.chat_format is None
and self.chat_handler is None
and "tokenizer.chat_template" in self.metadata
):
chat_format = llama_chat_format.guess_chat_format_from_gguf_metadata(
self.metadata
)
if chat_format is not None:
self.chat_format = chat_format
if self.verbose:
print(f"Guessed chat format: {chat_format}", file=sys.stderr)
else:
template = self.metadata["tokenizer.chat_template"]
try:
eos_token_id = int(self.metadata["tokenizer.ggml.eos_token_id"])
except:
eos_token_id = self.token_eos()
try:
bos_token_id = int(self.metadata["tokenizer.ggml.bos_token_id"])
except:
bos_token_id = self.token_bos()
eos_token = self._model.token_get_text(eos_token_id)
bos_token = self._model.token_get_text(bos_token_id)
if self.verbose:
print(f"Using gguf chat template: {template}", file=sys.stderr)
print(f"Using chat eos_token: {eos_token}", file=sys.stderr)
print(f"Using chat bos_token: {bos_token}", file=sys.stderr)
self.chat_handler = llama_chat_format.Jinja2ChatFormatter(
template=template,
eos_token=eos_token,
bos_token=bos_token,
stop_token_ids=[eos_token_id],
).to_chat_handler()
if self.chat_format is None and self.chat_handler is None:
self.chat_format = "llama-2"
if self.verbose:
print(f"Using fallback chat format: {chat_format}", file=sys.stderr)
2023-11-06 14:16:36 +00:00
@property
def ctx(self) -> llama_cpp.llama_context_p:
assert self._ctx.ctx is not None
return self._ctx.ctx
@property
def model(self) -> llama_cpp.llama_model_p:
assert self._model.model is not None
return self._model.model
@property
def _input_ids(self) -> npt.NDArray[np.intc]:
return self.input_ids[: self.n_tokens]
@property
def _scores(self) -> npt.NDArray[np.single]:
return self.scores[: self.n_tokens, :]
@property
def eval_tokens(self) -> Deque[int]:
return deque(self.input_ids[: self.n_tokens].tolist(), maxlen=self._n_ctx)
@property
def eval_logits(self) -> Deque[List[float]]:
return deque(
self.scores[: self.n_tokens, :].tolist(),
2023-09-30 20:02:35 +00:00
maxlen=self._n_ctx if self.context_params.logits_all else 1,
)
2023-05-26 20:12:45 +00:00
2023-11-06 14:16:36 +00:00
def tokenize(
self, text: bytes, add_bos: bool = True, special: bool = False
) -> List[int]:
"""Tokenize a string.
Args:
text: The utf-8 encoded string to tokenize.
2023-04-01 17:01:27 +00:00
Raises:
RuntimeError: If the tokenization failed.
Returns:
A list of tokens.
"""
return self.tokenizer_.tokenize(text, add_bos, special)
def detokenize(
self, tokens: List[int], prev_tokens: Optional[List[int]] = None
) -> bytes:
"""Detokenize a list of tokens.
Args:
tokens: The list of tokens to detokenize.
prev_tokens: The list of previous tokens. Offset mapping will be performed if provided
Returns:
The detokenized string.
"""
return self.tokenizer_.detokenize(tokens, prev_tokens=prev_tokens)
def set_cache(self, cache: Optional[BaseLlamaCache]):
2023-04-15 16:03:09 +00:00
"""Set the cache.
Args:
cache: The cache to set.
"""
self.cache = cache
2023-04-15 16:03:09 +00:00
2023-11-08 16:09:41 +00:00
def set_seed(self, seed: int):
"""Set the random seed.
Args:
seed: The random seed.
"""
assert self._ctx.ctx is not None
llama_cpp.llama_set_rng_seed(self._ctx.ctx, seed)
def reset(self):
"""Reset the model state."""
self.n_tokens = 0
def eval(self, tokens: Sequence[int]):
"""Evaluate a list of tokens.
Args:
tokens: The list of tokens to evaluate.
"""
2023-11-06 14:16:36 +00:00
assert self._ctx.ctx is not None
assert self._batch.batch is not None
2023-11-10 09:41:19 +00:00
self._ctx.kv_cache_seq_rm(-1, self.n_tokens, -1)
for i in range(0, len(tokens), self.n_batch):
batch = tokens[i : min(len(tokens), i + self.n_batch)]
2023-11-10 09:41:19 +00:00
n_past = self.n_tokens
n_tokens = len(batch)
2023-11-06 14:16:36 +00:00
self._batch.set_batch(
batch=batch, n_past=n_past, logits_all=self.context_params.logits_all
)
2023-11-06 14:16:36 +00:00
self._ctx.decode(self._batch)
2023-05-01 18:47:55 +00:00
# Save tokens
2023-11-10 10:15:41 +00:00
self.input_ids[n_past : n_past + n_tokens] = batch
2023-05-01 18:47:55 +00:00
# Save logits
if self.context_params.logits_all:
rows = n_tokens
cols = self._n_vocab
logits = self._ctx.get_logits()[: rows * cols]
self.scores[n_past : n_past + n_tokens, :].reshape(-1)[: :] = logits
else:
rows = 1
cols = self._n_vocab
logits = self._ctx.get_logits()[: rows * cols]
self.scores[n_past + n_tokens - 1, :].reshape(-1)[: :] = logits
# Update n_tokens
self.n_tokens += n_tokens
2023-05-01 18:47:55 +00:00
2023-11-06 14:16:36 +00:00
def sample(
self,
2023-11-06 14:16:36 +00:00
top_k: int = 40,
top_p: float = 0.95,
min_p: float = 0.05,
typical_p: float = 1.0,
2023-11-06 14:16:36 +00:00
temp: float = 0.80,
repeat_penalty: float = 1.1,
frequency_penalty: float = 0.0,
presence_penalty: float = 0.0,
tfs_z: float = 1.0,
mirostat_mode: int = 0,
mirostat_eta: float = 0.1,
mirostat_tau: float = 5.0,
penalize_nl: bool = True,
logits_processor: Optional[LogitsProcessorList] = None,
2023-08-08 19:08:54 +00:00
grammar: Optional[LlamaGrammar] = None,
idx: Optional[int] = None,
2023-05-01 18:47:55 +00:00
):
2023-11-06 14:16:36 +00:00
"""Sample a token from the model.
Args:
top_k: The top-k sampling parameter.
top_p: The top-p sampling parameter.
temp: The temperature parameter.
repeat_penalty: The repeat penalty parameter.
Returns:
The sampled token.
"""
assert self._ctx is not None
assert self.n_tokens > 0
if idx is None:
logits: npt.NDArray[np.single] = self._scores[-1, :]
else:
logits = self._scores[idx, :]
if logits_processor is not None:
logits[:] = (
logits_processor(self._input_ids, logits)
if idx is None
else logits_processor(self._input_ids[: idx + 1], logits)
)
sampling_params = _LlamaSamplingParams(
top_k=top_k,
top_p=top_p,
min_p=min_p,
tfs_z=tfs_z,
typical_p=typical_p,
temp=temp,
penalty_last_n=self.last_n_tokens_size,
2023-10-24 07:13:32 +00:00
penalty_repeat=repeat_penalty,
penalty_freq=frequency_penalty,
penalty_present=presence_penalty,
mirostat=mirostat_mode,
mirostat_tau=mirostat_tau,
mirostat_eta=mirostat_eta,
penalize_nl=penalize_nl,
)
sampling_context = _LlamaSamplingContext(
params=sampling_params,
grammar=grammar,
)
sampling_context.prev = list(self.eval_tokens)
id = sampling_context.sample(ctx_main=self._ctx, logits_array=logits)
sampling_context.accept(
ctx_main=self._ctx,
id=id,
apply_grammar=grammar is not None,
)
2023-08-06 17:21:37 +00:00
return id
2023-04-01 17:01:27 +00:00
def generate(
self,
tokens: Sequence[int],
top_k: int = 40,
top_p: float = 0.95,
min_p: float = 0.05,
typical_p: float = 1.0,
temp: float = 0.80,
repeat_penalty: float = 1.1,
reset: bool = True,
frequency_penalty: float = 0.0,
presence_penalty: float = 0.0,
tfs_z: float = 1.0,
mirostat_mode: int = 0,
mirostat_tau: float = 5.0,
mirostat_eta: float = 0.1,
penalize_nl: bool = True,
logits_processor: Optional[LogitsProcessorList] = None,
stopping_criteria: Optional[StoppingCriteriaList] = None,
2023-08-08 19:08:54 +00:00
grammar: Optional[LlamaGrammar] = None,
) -> Generator[int, Optional[Sequence[int]], None]:
"""Create a generator of tokens from a prompt.
2023-04-01 21:36:30 +00:00
2023-04-01 21:39:35 +00:00
Examples:
>>> llama = Llama("models/ggml-7b.bin")
>>> tokens = llama.tokenize(b"Hello, world!")
>>> for token in llama.generate(tokens, top_k=40, top_p=0.95, temp=1.0, repeat_penalty=1.1):
... print(llama.detokenize([token]))
2023-04-01 21:36:30 +00:00
Args:
tokens: The prompt tokens.
top_k: The top-k sampling parameter.
top_p: The top-p sampling parameter.
temp: The temperature parameter.
repeat_penalty: The repeat penalty parameter.
reset: Whether to reset the model state.
2023-04-01 21:36:30 +00:00
Yields:
The generated tokens.
"""
2024-01-19 13:31:59 +00:00
# Reset mirostat sampling
self._mirostat_mu = ctypes.c_float(2.0 * mirostat_tau)
# Check for kv cache prefix match
2023-11-06 14:16:36 +00:00
if reset and self.n_tokens > 0:
2023-05-05 01:58:27 +00:00
longest_prefix = 0
for a, b in zip(self._input_ids, tokens[:-1]):
2023-05-05 01:58:27 +00:00
if a == b:
longest_prefix += 1
else:
break
if longest_prefix > 0:
if self.verbose:
print("Llama.generate: prefix-match hit", file=sys.stderr)
reset = False
tokens = tokens[longest_prefix:]
self.n_tokens = longest_prefix
2024-01-19 13:31:59 +00:00
# Reset the model state
if reset:
self.reset()
2023-05-05 01:58:27 +00:00
2024-01-19 13:31:59 +00:00
# Reset the grammar
2023-08-08 19:08:54 +00:00
if grammar is not None:
grammar.reset()
2023-08-07 06:16:25 +00:00
sample_idx = self.n_tokens + len(tokens) - 1
tokens = list(tokens)
2024-01-19 13:31:59 +00:00
# Eval and sample
2023-04-01 17:01:27 +00:00
while True:
self.eval(tokens)
while sample_idx < self.n_tokens:
token = self.sample(
top_k=top_k,
top_p=top_p,
min_p=min_p,
typical_p=typical_p,
temp=temp,
repeat_penalty=repeat_penalty,
frequency_penalty=frequency_penalty,
presence_penalty=presence_penalty,
tfs_z=tfs_z,
mirostat_mode=mirostat_mode,
mirostat_tau=mirostat_tau,
mirostat_eta=mirostat_eta,
logits_processor=logits_processor,
grammar=grammar,
penalize_nl=penalize_nl,
idx=sample_idx,
)
sample_idx += 1
if stopping_criteria is not None and stopping_criteria(
self._input_ids, self._scores[-1, :]
):
return
tokens_or_none = yield token
tokens.clear()
tokens.append(token)
if tokens_or_none is not None:
tokens.extend(tokens_or_none)
if sample_idx < self.n_tokens and token != self._input_ids[sample_idx]:
self.n_tokens = sample_idx
self._ctx.kv_cache_seq_rm(-1, self.n_tokens, -1)
break
if self.draft_model is not None:
self.input_ids[self.n_tokens : self.n_tokens + len(tokens)] = tokens
draft_tokens = self.draft_model(
self.input_ids[: self.n_tokens + len(tokens)]
)
tokens.extend(
draft_tokens.astype(int)[
: self._n_ctx - self.n_tokens - len(tokens)
]
)
2023-04-01 17:01:27 +00:00
def create_embedding(
self, input: Union[str, List[str]], model: Optional[str] = None
2023-09-29 02:42:03 +00:00
) -> CreateEmbeddingResponse:
"""Embed a string.
Args:
2023-04-01 17:01:27 +00:00
input: The utf-8 encoded string to embed.
Returns:
2023-04-01 17:01:27 +00:00
An embedding object.
"""
2023-11-06 14:16:36 +00:00
assert self._model.model is not None
2023-05-16 22:07:25 +00:00
model_name: str = model if model is not None else self.model_path
2023-04-04 17:09:24 +00:00
input = input if isinstance(input, list) else [input]
# get numeric embeddings
embeds: List[List[float]]
total_tokens: int
embeds, total_tokens = self.embed(input, return_count=True) # type: ignore
# convert to CreateEmbeddingResponse
data: List[Embedding] = [
{
"object": "embedding",
"embedding": emb,
"index": idx,
}
for idx, emb in enumerate(embeds)
]
return {
"object": "list",
"data": data,
"model": model_name,
"usage": {
"prompt_tokens": total_tokens,
"total_tokens": total_tokens,
},
}
def embed(
self,
input: Union[str, List[str]],
normalize: bool = True,
truncate: bool = True,
return_count: bool = False,
):
"""Embed a string.
Args:
input: The utf-8 encoded string to embed.
Returns:
A list of embeddings
"""
assert self._ctx.ctx is not None
n_embd = self.n_embd()
n_batch = self.n_batch
2024-03-06 06:32:00 +00:00
if self.context_params.embeddings == False:
raise RuntimeError(
"Llama model must be created with embedding=True to call this method"
)
2023-04-04 17:09:24 +00:00
if self.verbose:
2023-11-06 14:16:36 +00:00
llama_cpp.llama_reset_timings(self._ctx.ctx)
2023-04-04 17:09:24 +00:00
if isinstance(input, str):
inputs = [input]
else:
inputs = input
2023-04-04 17:09:24 +00:00
# reset batch
self._batch.reset()
# decode and fetch embeddings
data: List[List[float]] = []
def decode_batch(n_seq: int):
assert self._ctx.ctx is not None
llama_cpp.llama_kv_cache_clear(self._ctx.ctx)
self._ctx.decode(self._batch)
self._batch.reset()
# store embeddings
for i in range(n_seq):
ptr = llama_cpp.llama_get_embeddings_seq(
self._ctx.ctx, i
)
if not ptr:
raise RuntimeError("Failed to get embeddings from sequence pooling type is not set")
embedding: List[float] = ptr[:n_embd]
if normalize:
norm = float(np.linalg.norm(embedding))
embedding = [v / norm for v in embedding]
data.append(embedding)
# init state
total_tokens = 0
t_batch = 0
p_batch = 0
# accumulate batches and encode
for text in inputs:
tokens = self.tokenize(text.encode("utf-8"))
if truncate:
tokens = tokens[:n_batch]
n_tokens = len(tokens)
total_tokens += n_tokens
# check for overrun
if n_tokens > n_batch:
raise ValueError(
f"Requested tokens ({n_tokens}) exceed batch size of {n_batch}"
)
# time to eval batch
if t_batch + n_tokens > n_batch:
decode_batch(p_batch)
t_batch = 0
p_batch = 0
# add to batch
self._batch.add_sequence(tokens, p_batch, False)
t_batch += n_tokens
p_batch += 1
# hanlde last batch
decode_batch(p_batch)
2023-05-22 01:30:03 +00:00
if self.verbose:
2023-11-06 14:16:36 +00:00
llama_cpp.llama_print_timings(self._ctx.ctx)
output = data[0] if isinstance(input, str) else data
2023-03-28 06:42:22 +00:00
llama_cpp.llama_kv_cache_clear(self._ctx.ctx)
self.reset()
2023-04-03 22:46:19 +00:00
if return_count:
return output, total_tokens
else:
return output
2023-04-03 22:46:19 +00:00
2023-04-01 17:01:27 +00:00
def _create_completion(
2023-03-23 09:33:06 +00:00
self,
prompt: Union[str, List[int]],
2023-03-23 09:33:06 +00:00
suffix: Optional[str] = None,
max_tokens: Optional[int] = 16,
2023-03-23 09:33:06 +00:00
temperature: float = 0.8,
top_p: float = 0.95,
min_p: float = 0.05,
typical_p: float = 1.0,
2023-03-23 19:51:05 +00:00
logprobs: Optional[int] = None,
2023-03-23 09:33:06 +00:00
echo: bool = False,
stop: Optional[Union[str, List[str]]] = [],
frequency_penalty: float = 0.0,
presence_penalty: float = 0.0,
2023-03-23 09:33:06 +00:00
repeat_penalty: float = 1.1,
top_k: int = 40,
stream: bool = False,
seed: Optional[int] = None,
tfs_z: float = 1.0,
mirostat_mode: int = 0,
mirostat_tau: float = 5.0,
mirostat_eta: float = 0.1,
model: Optional[str] = None,
stopping_criteria: Optional[StoppingCriteriaList] = None,
logits_processor: Optional[LogitsProcessorList] = None,
2023-08-08 19:08:54 +00:00
grammar: Optional[LlamaGrammar] = None,
logit_bias: Optional[Dict[str, float]] = None,
) -> Union[
Iterator[CreateCompletionResponse], Iterator[CreateCompletionStreamResponse]
]:
2023-11-06 14:16:36 +00:00
assert self._ctx is not None
assert suffix is None or suffix.__class__ is str
2023-05-24 20:02:06 +00:00
2023-04-15 15:39:21 +00:00
completion_id: str = f"cmpl-{str(uuid.uuid4())}"
created: int = int(time.time())
2023-11-21 03:50:59 +00:00
# If prompt is empty, initialize completion with BOS token to avoid
# detokenization including a space at the beginning of the completion
completion_tokens: List[int] = [] if len(prompt) > 0 else [self.token_bos()]
2023-04-01 17:01:27 +00:00
# Add blank space to start of prompt to match OG llama tokenizer
2023-09-29 02:42:03 +00:00
prompt_tokens: List[int] = (
2023-11-08 16:09:41 +00:00
(
self.tokenize(prompt.encode("utf-8"), special=True)
if prompt != ""
else [self.token_bos()]
)
if isinstance(prompt, str)
else prompt
)
2023-04-15 15:39:21 +00:00
text: bytes = b""
2023-05-18 15:35:59 +00:00
returned_tokens: int = 0
stop = (
stop if isinstance(stop, list) else [stop] if isinstance(stop, str) else []
)
2023-05-16 22:07:25 +00:00
model_name: str = model if model is not None else self.model_path
2023-03-23 09:33:06 +00:00
# NOTE: This likely doesn't work correctly for the first token in the prompt
# because of the extra space added to the start of the prompt_tokens
if logit_bias is not None:
logit_bias_map = {int(k): float(v) for k, v in logit_bias.items()}
def logit_bias_processor(
input_ids: npt.NDArray[np.intc],
scores: npt.NDArray[np.single],
) -> npt.NDArray[np.single]:
new_scores = np.copy(
scores
) # Does it make sense to copy the whole array or can we just overwrite the original one?
for input_id, score in logit_bias_map.items():
new_scores[input_id] = score + scores[input_id]
return new_scores
_logit_bias_processor = LogitsProcessorList([logit_bias_processor])
if logits_processor is None:
logits_processor = _logit_bias_processor
else:
logits_processor = logits_processor.extend(_logit_bias_processor)
2023-04-04 17:09:24 +00:00
if self.verbose:
2023-11-06 14:16:36 +00:00
self._ctx.reset_timings()
2023-04-04 17:09:24 +00:00
2023-11-06 14:16:36 +00:00
if len(prompt_tokens) >= self._n_ctx:
2023-03-23 09:33:06 +00:00
raise ValueError(
f"Requested tokens ({len(prompt_tokens)}) exceed context window of {llama_cpp.llama_n_ctx(self.ctx)}"
2023-03-23 09:33:06 +00:00
)
if max_tokens is None or max_tokens <= 0:
# Unlimited, depending on n_ctx.
2023-11-06 14:16:36 +00:00
max_tokens = self._n_ctx - len(prompt_tokens)
# Truncate max_tokens if requested tokens would exceed the context window
max_tokens = (
max_tokens
if max_tokens + len(prompt_tokens) < self._n_ctx
else (self._n_ctx - len(prompt_tokens))
)
2023-04-01 17:01:27 +00:00
if stop != []:
stop_sequences = [s.encode("utf-8") for s in stop]
2023-04-01 17:01:27 +00:00
else:
stop_sequences = []
2023-09-30 20:02:35 +00:00
if logprobs is not None and self.context_params.logits_all is False:
raise ValueError(
"logprobs is not supported for models created with logits_all=False"
)
2023-06-10 16:22:31 +00:00
if self.cache:
try:
cache_item = self.cache[prompt_tokens]
cache_prefix_len = Llama.longest_token_prefix(
cache_item.input_ids.tolist(), prompt_tokens
)
eval_prefix_len = Llama.longest_token_prefix(
self._input_ids.tolist(), prompt_tokens
)
if cache_prefix_len > eval_prefix_len:
self.load_state(cache_item)
if self.verbose:
print("Llama._create_completion: cache hit", file=sys.stderr)
except KeyError:
if self.verbose:
print("Llama._create_completion: cache miss", file=sys.stderr)
2023-11-08 16:09:41 +00:00
if seed is not None:
self._ctx.set_rng_seed(seed)
2023-04-15 16:03:09 +00:00
finish_reason = "length"
2023-04-28 10:50:30 +00:00
multibyte_fix = 0
2023-04-01 17:01:27 +00:00
for token in self.generate(
prompt_tokens,
top_k=top_k,
top_p=top_p,
min_p=min_p,
typical_p=typical_p,
2023-04-01 17:01:27 +00:00
temp=temperature,
tfs_z=tfs_z,
mirostat_mode=mirostat_mode,
mirostat_tau=mirostat_tau,
mirostat_eta=mirostat_eta,
frequency_penalty=frequency_penalty,
presence_penalty=presence_penalty,
2023-04-01 17:01:27 +00:00
repeat_penalty=repeat_penalty,
stopping_criteria=stopping_criteria,
logits_processor=logits_processor,
2023-08-08 19:08:54 +00:00
grammar=grammar,
):
assert self._model.model is not None
if llama_cpp.llama_token_is_eog(self._model.model, token):
text = self.detokenize(completion_tokens, prev_tokens=prompt_tokens)
2023-03-23 09:33:06 +00:00
finish_reason = "stop"
break
completion_tokens.append(token)
2023-03-23 09:33:06 +00:00
all_text = self.detokenize(completion_tokens, prev_tokens=prompt_tokens)
# Contains multi-byte UTF8
2023-05-01 18:47:55 +00:00
for k, char in enumerate(all_text[-3:]):
k = 3 - k
2023-05-01 18:47:55 +00:00
for num, pattern in [(2, 192), (3, 224), (4, 240)]:
# Bitwise AND check
2023-05-01 18:47:55 +00:00
if num > k and pattern & char == pattern:
multibyte_fix = num - k
2023-04-28 10:50:30 +00:00
# Stop incomplete bytes from passing
2023-05-01 18:47:55 +00:00
if multibyte_fix > 0:
2023-04-28 10:50:30 +00:00
multibyte_fix -= 1
continue
any_stop = [s for s in stop_sequences if s in all_text]
2023-03-23 09:33:06 +00:00
if len(any_stop) > 0:
first_stop = any_stop[0]
text = all_text[: all_text.index(first_stop)]
2023-03-23 09:33:06 +00:00
finish_reason = "stop"
break
if stream:
2023-05-27 00:23:49 +00:00
remaining_tokens = completion_tokens[returned_tokens:]
remaining_text = self.detokenize(remaining_tokens, prev_tokens=prompt_tokens + completion_tokens[:returned_tokens])
2023-05-27 00:23:49 +00:00
remaining_length = len(remaining_text)
# We want to avoid yielding any characters from
# the generated text if they are part of a stop
# sequence.
first_stop_position = 0
for s in stop_sequences:
2023-05-27 00:23:49 +00:00
for i in range(min(len(s), remaining_length), 0, -1):
if remaining_text.endswith(s[:i]):
if i > first_stop_position:
first_stop_position = i
break
2023-05-18 15:35:59 +00:00
token_end_position = 0
2023-08-09 14:04:35 +00:00
if logprobs is not None:
# not sure how to handle this branch when dealing
# with CJK output, so keep it unchanged
for token in remaining_tokens:
2023-11-21 03:50:59 +00:00
if token == self.token_bos():
continue
token_end_position += len(self.detokenize([token], prev_tokens=prompt_tokens + completion_tokens[:returned_tokens]))
2023-08-09 14:04:35 +00:00
# Check if stop sequence is in the token
2023-09-29 02:42:03 +00:00
if token_end_position > (
remaining_length - first_stop_position
):
2023-08-09 14:04:35 +00:00
break
token_str = self.detokenize([token], prev_tokens=prompt_tokens + completion_tokens[:returned_tokens]).decode(
"utf-8", errors="ignore"
)
text_offset = len(prompt) + len(
self.detokenize(completion_tokens[:returned_tokens], prev_tokens=prompt_tokens + completion_tokens[:returned_tokens]).decode(
"utf-8", errors="ignore"
)
)
token_offset = len(prompt_tokens) + returned_tokens
logits = self._scores[token_offset - 1, :]
current_logprobs = Llama.logits_to_logprobs(logits).tolist()
sorted_logprobs = list(
sorted(
zip(current_logprobs, range(len(current_logprobs))),
reverse=True,
)
)
top_logprob = {
self.detokenize([i]).decode(
"utf-8", errors="ignore"
): logprob
for logprob, i in sorted_logprobs[:logprobs]
}
top_logprob.update({token_str: current_logprobs[int(token)]})
logprobs_or_none = {
2024-02-09 07:02:13 +00:00
"tokens": [
self.detokenize([token], prev_tokens=prompt_tokens + completion_tokens[:returned_tokens]).decode(
2024-02-09 07:02:13 +00:00
"utf-8", errors="ignore"
)
],
"text_offset": [text_offset],
"token_logprobs": [current_logprobs[int(token)]],
"top_logprobs": [top_logprob],
}
2023-08-09 14:04:35 +00:00
returned_tokens += 1
yield {
"id": completion_id,
"object": "text_completion",
"created": created,
"model": model_name,
"choices": [
{
"text": self.detokenize([token], prev_tokens=prompt_tokens + completion_tokens[:returned_tokens]).decode(
2024-02-09 07:02:13 +00:00
"utf-8", errors="ignore"
),
2023-08-09 14:04:35 +00:00
"index": 0,
"logprobs": logprobs_or_none,
"finish_reason": None,
}
],
}
else:
while len(remaining_tokens) > 0:
decode_success = False
for i in range(1, len(remaining_tokens) + 1):
try:
bs = self.detokenize(remaining_tokens[:i], prev_tokens=prompt_tokens + completion_tokens[:returned_tokens])
2023-09-29 02:42:03 +00:00
ts = bs.decode("utf-8")
2023-08-09 14:04:35 +00:00
decode_success = True
break
except UnicodeError:
pass
2023-08-29 11:21:59 +00:00
else:
break
2023-08-09 14:04:35 +00:00
if not decode_success:
# all remaining tokens cannot be decoded to a UTF-8 character
break
token_end_position += len(bs)
2023-09-29 02:42:03 +00:00
if token_end_position > (
remaining_length - first_stop_position
):
2023-08-09 14:04:35 +00:00
break
remaining_tokens = remaining_tokens[i:]
returned_tokens += i
yield {
"id": completion_id,
"object": "text_completion",
"created": created,
"model": model_name,
"choices": [
{
2023-08-29 11:21:59 +00:00
"text": ts,
2023-08-09 14:04:35 +00:00
"index": 0,
"logprobs": None,
"finish_reason": None,
}
],
}
if len(completion_tokens) >= max_tokens:
text = self.detokenize(completion_tokens, prev_tokens=prompt_tokens)
finish_reason = "length"
break
2023-03-23 09:33:06 +00:00
if stopping_criteria is not None and stopping_criteria(
self._input_ids, self._scores[-1, :]
):
text = self.detokenize(completion_tokens, prev_tokens=prompt_tokens)
finish_reason = "stop"
if self.verbose:
2023-11-06 14:16:36 +00:00
self._ctx.print_timings()
if stream:
2023-05-18 15:35:59 +00:00
remaining_tokens = completion_tokens[returned_tokens:]
all_text = self.detokenize(remaining_tokens, prev_tokens=prompt_tokens + completion_tokens[:returned_tokens])
2023-05-18 15:35:59 +00:00
any_stop = [s for s in stop_sequences if s in all_text]
if len(any_stop) > 0:
end = min(all_text.index(stop) for stop in any_stop)
else:
end = len(all_text)
token_end_position = 0
2023-05-18 15:35:59 +00:00
for token in remaining_tokens:
token_end_position += len(self.detokenize([token], prev_tokens=prompt_tokens + completion_tokens[:returned_tokens]))
logprobs_or_none: Optional[CompletionLogprobs] = None
if logprobs is not None:
2023-11-21 03:50:59 +00:00
if token == self.token_bos():
continue
token_str = self.detokenize([token]).decode(
"utf-8", errors="ignore"
)
text_offset = len(prompt) + len(
self.detokenize(completion_tokens[:returned_tokens], prev_tokens=prompt_tokens + completion_tokens[:returned_tokens])
)
token_offset = len(prompt_tokens) + returned_tokens - 1
logits = self._scores[token_offset, :]
current_logprobs = Llama.logits_to_logprobs(logits).tolist()
sorted_logprobs = list(
sorted(
zip(current_logprobs, range(len(current_logprobs))),
reverse=True,
)
)
top_logprob = {
self.detokenize([i]).decode("utf-8", errors="ignore"): logprob
for logprob, i in sorted_logprobs[:logprobs]
}
top_logprob.update({token_str: current_logprobs[int(token)]})
logprobs_or_none = {
"tokens": [
self.detokenize([token]).decode("utf-8", errors="ignore")
],
"text_offset": [text_offset],
"token_logprobs": [current_logprobs[int(token)]],
"top_logprobs": [top_logprob],
}
if token_end_position >= end:
2023-05-18 15:35:59 +00:00
last_text = self.detokenize([token])
if token_end_position == end - 1:
2023-05-18 15:35:59 +00:00
break
returned_tokens += 1
2023-05-18 15:35:59 +00:00
yield {
"id": completion_id,
"object": "text_completion",
"created": created,
"model": model_name,
"choices": [
{
"text": last_text[
: len(last_text) - (token_end_position - end)
].decode("utf-8", errors="ignore"),
2023-05-18 15:35:59 +00:00
"index": 0,
"logprobs": logprobs_or_none,
"finish_reason": None,
}
],
}
2023-05-18 15:35:59 +00:00
break
returned_tokens += 1
yield {
"id": completion_id,
"object": "text_completion",
"created": created,
2023-05-18 15:35:59 +00:00
"model": model_name,
"choices": [
{
2023-05-18 15:35:59 +00:00
"text": self.detokenize([token]).decode(
"utf-8", errors="ignore"
),
"index": 0,
"logprobs": logprobs_or_none,
"finish_reason": None,
}
],
}
yield {
"id": completion_id,
"object": "text_completion",
"created": created,
"model": model_name,
"choices": [
{
"text": "",
"index": 0,
"logprobs": None,
"finish_reason": finish_reason,
}
],
}
2023-06-10 16:22:31 +00:00
if self.cache:
if self.verbose:
print("Llama._create_completion: cache save", file=sys.stderr)
self.cache[prompt_tokens + completion_tokens] = self.save_state()
print("Llama._create_completion: cache saved", file=sys.stderr)
return
2023-06-10 16:22:31 +00:00
if self.cache:
if self.verbose:
print("Llama._create_completion: cache save", file=sys.stderr)
self.cache[prompt_tokens + completion_tokens] = self.save_state()
2023-04-26 12:37:06 +00:00
text_str = text.decode("utf-8", errors="ignore")
2023-03-23 09:33:06 +00:00
if echo:
2023-04-15 16:03:09 +00:00
text_str = prompt + text_str
2023-03-23 09:33:06 +00:00
if suffix is not None:
2023-04-15 16:03:09 +00:00
text_str = text_str + suffix
2023-03-23 09:33:06 +00:00
logprobs_or_none: Optional[CompletionLogprobs] = None
2023-03-23 19:51:05 +00:00
if logprobs is not None:
text_offset = 0 if echo else len(prompt)
token_offset = 0 if echo else len(prompt_tokens[1:])
2023-04-14 13:59:33 +00:00
text_offsets: List[int] = []
token_logprobs: List[Optional[float]] = []
2023-04-14 13:59:33 +00:00
tokens: List[str] = []
top_logprobs: List[Optional[Dict[str, float]]] = []
if echo:
# Remove leading BOS token
all_tokens = prompt_tokens[1:] + completion_tokens
else:
all_tokens = completion_tokens
2023-04-14 13:59:33 +00:00
all_token_strs = [
self.detokenize([token], prev_tokens=all_tokens[:i]).decode("utf-8", errors="ignore")
for i, token in enumerate(all_tokens)
2023-04-14 13:59:33 +00:00
]
all_logprobs = Llama.logits_to_logprobs(self._scores)[token_offset:]
# TODO: may be able to change this loop to use np.take_along_dim
for idx, (token, token_str, logprobs_token) in enumerate(
zip(all_tokens, all_token_strs, all_logprobs)
2023-04-14 13:59:33 +00:00
):
2023-11-21 03:50:59 +00:00
if token == self.token_bos():
continue
text_offsets.append(
text_offset
+ len(
self.detokenize(all_tokens[:idx]).decode(
"utf-8", errors="ignore"
)
)
)
2023-04-14 13:59:33 +00:00
tokens.append(token_str)
sorted_logprobs = list(
sorted(
zip(logprobs_token, range(len(logprobs_token))), reverse=True
)
)
token_logprobs.append(logprobs_token[int(token)])
top_logprob: Optional[Dict[str, float]] = {
self.detokenize([i], prev_tokens=all_tokens[:idx]).decode("utf-8", errors="ignore"): logprob
2023-04-14 13:59:33 +00:00
for logprob, i in sorted_logprobs[:logprobs]
}
top_logprob.update({token_str: logprobs_token[int(token)]})
2023-04-14 13:59:33 +00:00
top_logprobs.append(top_logprob)
# Weird idosincracy of the OpenAI API where
# token_logprobs and top_logprobs are null for
# the first token.
if echo and len(all_tokens) > 0:
token_logprobs[0] = None
top_logprobs[0] = None
logprobs_or_none = {
"tokens": tokens,
"text_offset": text_offsets,
"token_logprobs": token_logprobs,
"top_logprobs": top_logprobs,
}
2023-04-04 17:09:24 +00:00
yield {
2023-03-28 06:42:22 +00:00
"id": completion_id,
2023-03-23 09:33:06 +00:00
"object": "text_completion",
2023-03-28 06:42:22 +00:00
"created": created,
2023-05-16 22:07:25 +00:00
"model": model_name,
2023-03-23 09:33:06 +00:00
"choices": [
{
2023-04-15 16:03:09 +00:00
"text": text_str,
2023-03-23 09:33:06 +00:00
"index": 0,
"logprobs": logprobs_or_none,
2023-03-23 09:33:06 +00:00
"finish_reason": finish_reason,
}
],
"usage": {
"prompt_tokens": len(prompt_tokens),
"completion_tokens": len(completion_tokens),
"total_tokens": len(prompt_tokens) + len(completion_tokens),
2023-03-23 09:33:06 +00:00
},
}
2023-04-01 17:01:27 +00:00
def create_completion(
self,
prompt: Union[str, List[int]],
2023-04-01 17:01:27 +00:00
suffix: Optional[str] = None,
max_tokens: Optional[int] = 16,
2023-04-01 17:01:27 +00:00
temperature: float = 0.8,
top_p: float = 0.95,
min_p: float = 0.05,
typical_p: float = 1.0,
2023-04-01 17:01:27 +00:00
logprobs: Optional[int] = None,
echo: bool = False,
stop: Optional[Union[str, List[str]]] = [],
frequency_penalty: float = 0.0,
presence_penalty: float = 0.0,
2023-04-01 17:01:27 +00:00
repeat_penalty: float = 1.1,
top_k: int = 40,
stream: bool = False,
seed: Optional[int] = None,
tfs_z: float = 1.0,
mirostat_mode: int = 0,
mirostat_tau: float = 5.0,
mirostat_eta: float = 0.1,
model: Optional[str] = None,
stopping_criteria: Optional[StoppingCriteriaList] = None,
logits_processor: Optional[LogitsProcessorList] = None,
2023-08-08 19:08:54 +00:00
grammar: Optional[LlamaGrammar] = None,
logit_bias: Optional[Dict[str, float]] = None,
) -> Union[CreateCompletionResponse, Iterator[CreateCompletionStreamResponse]]:
2023-04-01 17:01:27 +00:00
"""Generate text from a prompt.
Args:
prompt: The prompt to generate text from.
suffix: A suffix to append to the generated text. If None, no suffix is appended.
max_tokens: The maximum number of tokens to generate. If max_tokens <= 0 or None, the maximum number of tokens to generate is unlimited and depends on n_ctx.
2023-04-01 17:01:27 +00:00
temperature: The temperature to use for sampling.
top_p: The top-p value to use for nucleus sampling. Nucleus sampling described in academic paper "The Curious Case of Neural Text Degeneration" https://arxiv.org/abs/1904.09751
min_p: The min-p value to use for minimum p sampling. Minimum P sampling as described in https://github.com/ggerganov/llama.cpp/pull/3841
typical_p: The typical-p value to use for sampling. Locally Typical Sampling implementation described in the paper https://arxiv.org/abs/2202.00666.
2023-04-01 17:01:27 +00:00
logprobs: The number of logprobs to return. If None, no logprobs are returned.
echo: Whether to echo the prompt.
stop: A list of strings to stop generation when encountered.
frequency_penalty: The penalty to apply to tokens based on their frequency in the prompt.
presence_penalty: The penalty to apply to tokens based on their presence in the prompt.
2023-04-01 17:01:27 +00:00
repeat_penalty: The penalty to apply to repeated tokens.
top_k: The top-k value to use for sampling. Top-K sampling described in academic paper "The Curious Case of Neural Text Degeneration" https://arxiv.org/abs/1904.09751
2023-04-01 17:01:27 +00:00
stream: Whether to stream the results.
seed: The seed to use for sampling.
tfs_z: The tail-free sampling parameter. Tail Free Sampling described in https://www.trentonbricken.com/Tail-Free-Sampling/.
mirostat_mode: The mirostat sampling mode.
mirostat_tau: The target cross-entropy (or surprise) value you want to achieve for the generated text. A higher value corresponds to more surprising or less predictable text, while a lower value corresponds to less surprising or more predictable text.
mirostat_eta: The learning rate used to update `mu` based on the error between the target and observed surprisal of the sampled word. A larger learning rate will cause `mu` to be updated more quickly, while a smaller learning rate will result in slower updates.
model: The name to use for the model in the completion object.
stopping_criteria: A list of stopping criteria to use.
logits_processor: A list of logits processors to use.
grammar: A grammar to use for constrained sampling.
logit_bias: A logit bias to use.
2023-04-01 17:01:27 +00:00
Raises:
ValueError: If the requested tokens exceed the context window.
RuntimeError: If the prompt fails to tokenize or the model fails to evaluate the prompt.
Returns:
Response object containing the generated text.
"""
completion_or_chunks = self._create_completion(
prompt=prompt,
suffix=suffix,
max_tokens=-1 if max_tokens is None else max_tokens,
2023-04-01 17:01:27 +00:00
temperature=temperature,
top_p=top_p,
min_p=min_p,
typical_p=typical_p,
2023-04-01 17:01:27 +00:00
logprobs=logprobs,
echo=echo,
stop=stop,
frequency_penalty=frequency_penalty,
presence_penalty=presence_penalty,
2023-04-01 17:01:27 +00:00
repeat_penalty=repeat_penalty,
top_k=top_k,
stream=stream,
seed=seed,
2023-05-12 01:56:19 +00:00
tfs_z=tfs_z,
mirostat_mode=mirostat_mode,
mirostat_tau=mirostat_tau,
mirostat_eta=mirostat_eta,
model=model,
stopping_criteria=stopping_criteria,
logits_processor=logits_processor,
2023-09-29 02:42:03 +00:00
grammar=grammar,
logit_bias=logit_bias,
2023-04-01 17:01:27 +00:00
)
if stream:
chunks: Iterator[CreateCompletionStreamResponse] = completion_or_chunks
2023-04-01 17:01:27 +00:00
return chunks
completion: Completion = next(completion_or_chunks) # type: ignore
return completion
def __call__(
self,
prompt: str,
suffix: Optional[str] = None,
max_tokens: Optional[int] = 16,
temperature: float = 0.8,
top_p: float = 0.95,
min_p: float = 0.05,
typical_p: float = 1.0,
logprobs: Optional[int] = None,
echo: bool = False,
stop: Optional[Union[str, List[str]]] = [],
frequency_penalty: float = 0.0,
presence_penalty: float = 0.0,
repeat_penalty: float = 1.1,
top_k: int = 40,
stream: bool = False,
seed: Optional[int] = None,
tfs_z: float = 1.0,
mirostat_mode: int = 0,
mirostat_tau: float = 5.0,
mirostat_eta: float = 0.1,
model: Optional[str] = None,
stopping_criteria: Optional[StoppingCriteriaList] = None,
logits_processor: Optional[LogitsProcessorList] = None,
2023-08-08 19:08:54 +00:00
grammar: Optional[LlamaGrammar] = None,
logit_bias: Optional[Dict[str, float]] = None,
) -> Union[CreateCompletionResponse, Iterator[CreateCompletionStreamResponse]]:
"""Generate text from a prompt.
Args:
prompt: The prompt to generate text from.
suffix: A suffix to append to the generated text. If None, no suffix is appended.
max_tokens: The maximum number of tokens to generate. If max_tokens <= 0 or None, the maximum number of tokens to generate is unlimited and depends on n_ctx.
temperature: The temperature to use for sampling.
top_p: The top-p value to use for nucleus sampling. Nucleus sampling described in academic paper "The Curious Case of Neural Text Degeneration" https://arxiv.org/abs/1904.09751
min_p: The min-p value to use for minimum p sampling. Minimum P sampling as described in https://github.com/ggerganov/llama.cpp/pull/3841
typical_p: The typical-p value to use for sampling. Locally Typical Sampling implementation described in the paper https://arxiv.org/abs/2202.00666.
logprobs: The number of logprobs to return. If None, no logprobs are returned.
echo: Whether to echo the prompt.
stop: A list of strings to stop generation when encountered.
frequency_penalty: The penalty to apply to tokens based on their frequency in the prompt.
presence_penalty: The penalty to apply to tokens based on their presence in the prompt.
repeat_penalty: The penalty to apply to repeated tokens.
top_k: The top-k value to use for sampling. Top-K sampling described in academic paper "The Curious Case of Neural Text Degeneration" https://arxiv.org/abs/1904.09751
stream: Whether to stream the results.
seed: The seed to use for sampling.
tfs_z: The tail-free sampling parameter. Tail Free Sampling described in https://www.trentonbricken.com/Tail-Free-Sampling/.
mirostat_mode: The mirostat sampling mode.
mirostat_tau: The target cross-entropy (or surprise) value you want to achieve for the generated text. A higher value corresponds to more surprising or less predictable text, while a lower value corresponds to less surprising or more predictable text.
mirostat_eta: The learning rate used to update `mu` based on the error between the target and observed surprisal of the sampled word. A larger learning rate will cause `mu` to be updated more quickly, while a smaller learning rate will result in slower updates.
model: The name to use for the model in the completion object.
stopping_criteria: A list of stopping criteria to use.
logits_processor: A list of logits processors to use.
grammar: A grammar to use for constrained sampling.
logit_bias: A logit bias to use.
Raises:
ValueError: If the requested tokens exceed the context window.
RuntimeError: If the prompt fails to tokenize or the model fails to evaluate the prompt.
Returns:
Response object containing the generated text.
"""
2023-04-01 17:01:27 +00:00
return self.create_completion(
prompt=prompt,
suffix=suffix,
max_tokens=max_tokens,
temperature=temperature,
top_p=top_p,
min_p=min_p,
typical_p=typical_p,
logprobs=logprobs,
echo=echo,
stop=stop,
frequency_penalty=frequency_penalty,
presence_penalty=presence_penalty,
repeat_penalty=repeat_penalty,
top_k=top_k,
stream=stream,
seed=seed,
2023-05-12 01:56:19 +00:00
tfs_z=tfs_z,
mirostat_mode=mirostat_mode,
mirostat_tau=mirostat_tau,
mirostat_eta=mirostat_eta,
model=model,
stopping_criteria=stopping_criteria,
logits_processor=logits_processor,
2023-08-08 19:08:54 +00:00
grammar=grammar,
logit_bias=logit_bias,
)
2023-04-04 00:12:44 +00:00
def create_chat_completion(
self,
messages: List[ChatCompletionRequestMessage],
2023-07-19 07:48:20 +00:00
functions: Optional[List[ChatCompletionFunction]] = None,
function_call: Optional[ChatCompletionRequestFunctionCall] = None,
tools: Optional[List[ChatCompletionTool]] = None,
tool_choice: Optional[ChatCompletionToolChoiceOption] = None,
temperature: float = 0.2,
2023-04-04 00:12:44 +00:00
top_p: float = 0.95,
top_k: int = 40,
min_p: float = 0.05,
typical_p: float = 1.0,
2023-04-04 00:12:44 +00:00
stream: bool = False,
stop: Optional[Union[str, List[str]]] = [],
seed: Optional[int] = None,
2023-11-08 05:07:16 +00:00
response_format: Optional[ChatCompletionRequestResponseFormat] = None,
max_tokens: Optional[int] = None,
presence_penalty: float = 0.0,
frequency_penalty: float = 0.0,
2023-04-04 00:12:44 +00:00
repeat_penalty: float = 1.1,
tfs_z: float = 1.0,
mirostat_mode: int = 0,
mirostat_tau: float = 5.0,
mirostat_eta: float = 0.1,
model: Optional[str] = None,
logits_processor: Optional[LogitsProcessorList] = None,
2023-08-08 19:08:54 +00:00
grammar: Optional[LlamaGrammar] = None,
logit_bias: Optional[Dict[str, float]] = None,
logprobs: Optional[bool] = None,
top_logprobs: Optional[int] = None,
) -> Union[
CreateChatCompletionResponse, Iterator[CreateChatCompletionStreamResponse]
]:
"""Generate a chat completion from a list of messages.
Args:
messages: A list of messages to generate a response for.
functions: A list of functions to use for the chat completion.
function_call: A function call to use for the chat completion.
tools: A list of tools to use for the chat completion.
tool_choice: A tool choice to use for the chat completion.
temperature: The temperature to use for sampling.
top_p: The top-p value to use for nucleus sampling. Nucleus sampling described in academic paper "The Curious Case of Neural Text Degeneration" https://arxiv.org/abs/1904.09751
top_k: The top-k value to use for sampling. Top-K sampling described in academic paper "The Curious Case of Neural Text Degeneration" https://arxiv.org/abs/1904.09751
min_p: The min-p value to use for minimum p sampling. Minimum P sampling as described in https://github.com/ggerganov/llama.cpp/pull/3841
typical_p: The typical-p value to use for sampling. Locally Typical Sampling implementation described in the paper https://arxiv.org/abs/2202.00666.
stream: Whether to stream the results.
stop: A list of strings to stop generation when encountered.
seed: The seed to use for sampling.
response_format: The response format to use for the chat completion. Use { "type": "json_object" } to contstrain output to only valid json.
max_tokens: The maximum number of tokens to generate. If max_tokens <= 0 or None, the maximum number of tokens to generate is unlimited and depends on n_ctx.
presence_penalty: The penalty to apply to tokens based on their presence in the prompt.
frequency_penalty: The penalty to apply to tokens based on their frequency in the prompt.
repeat_penalty: The penalty to apply to repeated tokens.
tfs_z: The tail-free sampling parameter.
mirostat_mode: The mirostat sampling mode.
mirostat_tau: The mirostat sampling tau parameter.
mirostat_eta: The mirostat sampling eta parameter.
model: The name to use for the model in the completion object.
logits_processor: A list of logits processors to use.
grammar: A grammar to use.
logit_bias: A logit bias to use.
Returns:
Generated chat completion or a stream of chat completion chunks.
"""
handler = self.chat_handler or llama_chat_format.get_chat_completion_handler(
self.chat_format
)
return handler(
llama=self,
messages=messages,
functions=functions,
function_call=function_call,
tools=tools,
tool_choice=tool_choice,
2023-04-04 00:12:44 +00:00
temperature=temperature,
top_p=top_p,
top_k=top_k,
min_p=min_p,
typical_p=typical_p,
logprobs=logprobs,
top_logprobs=top_logprobs,
2023-04-04 00:12:44 +00:00
stream=stream,
stop=stop,
seed=seed,
2023-11-08 05:07:16 +00:00
response_format=response_format,
2023-04-04 00:12:44 +00:00
max_tokens=max_tokens,
presence_penalty=presence_penalty,
frequency_penalty=frequency_penalty,
repeat_penalty=repeat_penalty,
2023-05-12 01:56:19 +00:00
tfs_z=tfs_z,
mirostat_mode=mirostat_mode,
mirostat_tau=mirostat_tau,
mirostat_eta=mirostat_eta,
model=model,
logits_processor=logits_processor,
2023-08-08 19:08:54 +00:00
grammar=grammar,
logit_bias=logit_bias,
2023-04-04 00:12:44 +00:00
)
def create_chat_completion_openai_v1(
self,
*args: Any,
**kwargs: Any,
):
"""Generate a chat completion with return type based on the the OpenAI v1 API.
OpenAI python package is required to use this method.
You can install it with `pip install openai`.
Args:
*args: Positional arguments to pass to create_chat_completion.
**kwargs: Keyword arguments to pass to create_chat_completion.
Returns:
Generated chat completion or a stream of chat completion chunks.
"""
try:
from openai.types.chat import ChatCompletion, ChatCompletionChunk
stream = kwargs.get("stream", False) # type: ignore
assert isinstance(stream, bool)
if stream:
return (ChatCompletionChunk(**chunk) for chunk in self.create_chat_completion(*args, **kwargs)) # type: ignore
else:
return ChatCompletion(**self.create_chat_completion(*args, **kwargs)) # type: ignore
except ImportError:
raise ImportError(
"To use create_chat_completion_openai_v1, you must install the openai package."
"You can install it with `pip install openai`."
)
def __getstate__(self):
return dict(
model_path=self.model_path,
2023-09-29 02:42:03 +00:00
# Model Params
n_gpu_layers=self.model_params.n_gpu_layers,
2024-01-15 17:49:20 +00:00
split_mode=self.model_params.split_mode,
2023-09-29 02:42:03 +00:00
main_gpu=self.model_params.main_gpu,
tensor_split=self.tensor_split,
vocab_only=self.model_params.vocab_only,
use_mmap=self.model_params.use_mmap,
use_mlock=self.model_params.use_mlock,
kv_overrides=self.kv_overrides,
2023-09-29 02:42:03 +00:00
# Context Params
seed=self.context_params.seed,
n_ctx=self.context_params.n_ctx,
n_batch=self.n_batch,
2023-09-29 02:42:03 +00:00
n_threads=self.context_params.n_threads,
n_threads_batch=self.context_params.n_threads_batch,
rope_scaling_type=self.context_params.rope_scaling_type,
pooling_type=self.context_params.pooling_type,
2023-09-29 02:42:03 +00:00
rope_freq_base=self.context_params.rope_freq_base,
rope_freq_scale=self.context_params.rope_freq_scale,
yarn_ext_factor=self.context_params.yarn_ext_factor,
yarn_attn_factor=self.context_params.yarn_attn_factor,
yarn_beta_fast=self.context_params.yarn_beta_fast,
yarn_beta_slow=self.context_params.yarn_beta_slow,
yarn_orig_ctx=self.context_params.yarn_orig_ctx,
2023-09-29 02:42:03 +00:00
logits_all=self.context_params.logits_all,
2024-03-06 06:32:00 +00:00
embedding=self.context_params.embeddings,
offload_kqv=self.context_params.offload_kqv,
2023-09-29 02:42:03 +00:00
# Sampling Params
last_n_tokens_size=self.last_n_tokens_size,
# LoRA Params
lora_base=self.lora_base,
2023-09-29 02:42:03 +00:00
lora_scale=self.lora_scale,
2023-04-18 05:43:44 +00:00
lora_path=self.lora_path,
2023-09-29 02:42:03 +00:00
# Backend Params
numa=self.numa,
# Chat Format Params
chat_format=self.chat_format,
chat_handler=self.chat_handler,
# Speculative Decidng
draft_model=self.draft_model,
# KV cache quantization
type_k=self.context_params.type_k,
type_v=self.context_params.type_v,
2023-09-29 02:42:03 +00:00
# Misc
verbose=self.verbose,
)
def __setstate__(self, state):
self.__init__(**state)
def save_state(self) -> LlamaState:
2023-11-06 14:16:36 +00:00
assert self._ctx.ctx is not None
if self.verbose:
print("Llama.save_state: saving llama state", file=sys.stderr)
2023-11-06 14:16:36 +00:00
state_size = llama_cpp.llama_get_state_size(self._ctx.ctx)
if self.verbose:
print(f"Llama.save_state: got state size: {state_size}", file=sys.stderr)
llama_state = (ctypes.c_uint8 * int(state_size))()
if self.verbose:
print("Llama.save_state: allocated state", file=sys.stderr)
2023-11-06 14:16:36 +00:00
n_bytes = llama_cpp.llama_copy_state_data(self._ctx.ctx, llama_state)
if self.verbose:
print(f"Llama.save_state: copied llama state: {n_bytes}", file=sys.stderr)
2023-05-03 13:33:50 +00:00
if int(n_bytes) > int(state_size):
raise RuntimeError("Failed to copy llama state data")
llama_state_compact = (ctypes.c_uint8 * int(n_bytes))()
2023-05-03 13:33:50 +00:00
llama_cpp.ctypes.memmove(llama_state_compact, llama_state, int(n_bytes))
2023-05-03 14:28:10 +00:00
if self.verbose:
2023-05-05 01:58:36 +00:00
print(
f"Llama.save_state: saving {n_bytes} bytes of llama state",
file=sys.stderr,
)
return LlamaState(
scores=self._scores.copy(),
input_ids=self.input_ids.copy(),
n_tokens=self.n_tokens,
llama_state=bytes(llama_state_compact),
2023-05-03 13:33:50 +00:00
llama_state_size=n_bytes,
)
def load_state(self, state: LlamaState) -> None:
2023-11-06 14:16:36 +00:00
assert self._ctx.ctx is not None
# Only filling in up to `n_tokens` and then zero-ing out the rest
self.scores[: state.n_tokens, :] = state.scores.copy()
self.scores[state.n_tokens :, :] = 0.0
self.input_ids = state.input_ids.copy()
self.n_tokens = state.n_tokens
2023-05-03 13:33:50 +00:00
state_size = state.llama_state_size
LLamaStateArrayType = ctypes.c_uint8 * state_size
llama_state = LLamaStateArrayType.from_buffer_copy(state.llama_state)
2023-11-06 14:16:36 +00:00
if llama_cpp.llama_set_state_data(self._ctx.ctx, llama_state) != state_size:
raise RuntimeError("Failed to set llama state data")
def n_ctx(self) -> int:
"""Return the context window size."""
2023-11-06 14:16:36 +00:00
return self._ctx.n_ctx()
def n_embd(self) -> int:
"""Return the embedding size."""
2023-11-06 14:16:36 +00:00
return self._model.n_embd()
def n_vocab(self) -> int:
"""Return the vocabulary size."""
2023-11-06 14:16:36 +00:00
return self._model.n_vocab()
def tokenizer(self) -> LlamaTokenizer:
"""Return the llama tokenizer for this model."""
2023-05-25 18:11:33 +00:00
return LlamaTokenizer(self)
2023-08-24 04:17:00 +00:00
def token_eos(self) -> int:
"""Return the end-of-sequence token."""
2023-11-06 14:16:36 +00:00
return self._model.token_eos()
2023-08-24 04:17:00 +00:00
def token_bos(self) -> int:
"""Return the beginning-of-sequence token."""
2023-11-06 14:16:36 +00:00
return self._model.token_bos()
2023-08-24 04:17:00 +00:00
def token_nl(self) -> int:
2023-05-17 05:53:26 +00:00
"""Return the newline token."""
2023-11-06 14:16:36 +00:00
return self._model.token_nl()
2023-05-17 05:53:26 +00:00
@staticmethod
def logits_to_logprobs(
logits: Union[npt.NDArray[np.single], List], axis: int = -1
) -> npt.NDArray[np.single]:
# https://docs.scipy.org/doc/scipy/reference/generated/scipy.special.log_softmax.html
logits_maxs: np.ndarray = np.amax(logits, axis=axis, keepdims=True)
if logits_maxs.ndim > 0:
logits_maxs[~np.isfinite(logits_maxs)] = 0
elif not np.isfinite(logits_maxs):
logits_maxs = 0
subtract_maxs = np.subtract(logits, logits_maxs, dtype=np.single)
exp = np.exp(subtract_maxs)
# Suppress warnings about log of zero
with np.errstate(divide="ignore"):
summed = np.sum(exp, axis=axis, keepdims=True)
out = np.log(summed)
return subtract_maxs - out
@staticmethod
def longest_token_prefix(a: Sequence[int], b: Sequence[int]):
longest_prefix = 0
for _a, _b in zip(a, b):
if _a == _b:
longest_prefix += 1
else:
break
return longest_prefix
2023-05-25 18:11:33 +00:00
@classmethod
def from_pretrained(
cls,
repo_id: str,
filename: Optional[str],
local_dir: Optional[Union[str, os.PathLike[str]]] = None,
local_dir_use_symlinks: Union[bool, Literal["auto"]] = "auto",
cache_dir: Optional[Union[str, os.PathLike[str]]] = None,
**kwargs: Any,
) -> "Llama":
"""Create a Llama model from a pretrained model name or path.
This method requires the huggingface-hub package.
You can install it with `pip install huggingface-hub`.
Args:
repo_id: The model repo id.
filename: A filename or glob pattern to match the model file in the repo.
local_dir: The local directory to save the model to.
local_dir_use_symlinks: Whether to use symlinks when downloading the model.
**kwargs: Additional keyword arguments to pass to the Llama constructor.
Returns:
A Llama model."""
try:
from huggingface_hub import hf_hub_download, HfFileSystem
from huggingface_hub.utils import validate_repo_id
except ImportError:
raise ImportError(
"Llama.from_pretrained requires the huggingface-hub package. "
"You can install it with `pip install huggingface-hub`."
)
validate_repo_id(repo_id)
hffs = HfFileSystem()
files = [
file["name"] if isinstance(file, dict) else file
for file in hffs.ls(repo_id)
]
# split each file into repo_id, subfolder, filename
file_list: List[str] = []
for file in files:
rel_path = Path(file).relative_to(repo_id)
file_list.append(str(rel_path))
2023-05-25 18:11:33 +00:00
matching_files = [file for file in file_list if fnmatch.fnmatch(file, filename)] # type: ignore
if len(matching_files) == 0:
raise ValueError(
f"No file found in {repo_id} that match {filename}\n\n"
f"Available Files:\n{json.dumps(file_list)}"
)
if len(matching_files) > 1:
raise ValueError(
f"Multiple files found in {repo_id} matching {filename}\n\n"
f"Available Files:\n{json.dumps(files)}"
)
(matching_file,) = matching_files
subfolder = str(Path(matching_file).parent)
filename = Path(matching_file).name
# download the file
hf_hub_download(
repo_id=repo_id,
filename=filename,
subfolder=subfolder,
local_dir=local_dir,
local_dir_use_symlinks=local_dir_use_symlinks,
cache_dir=cache_dir,
)
if local_dir is None:
model_path = hf_hub_download(
repo_id=repo_id,
filename=filename,
subfolder=subfolder,
local_dir=local_dir,
local_dir_use_symlinks=local_dir_use_symlinks,
cache_dir=cache_dir,
local_files_only=True,
)
else:
model_path = os.path.join(local_dir, filename)
return cls(
model_path=model_path,
**kwargs,
)
2024-01-17 14:16:13 +00:00
class LlamaState:
def __init__(
self,
input_ids: npt.NDArray[np.intc],
scores: npt.NDArray[np.single],
n_tokens: int,
llama_state: bytes,
llama_state_size: int,
):
self.input_ids = input_ids
self.scores = scores
self.n_tokens = n_tokens
self.llama_state = llama_state
self.llama_state_size = llama_state_size
LogitsProcessor = Callable[
[npt.NDArray[np.intc], npt.NDArray[np.single]], npt.NDArray[np.single]
]
class LogitsProcessorList(List[LogitsProcessor]):
def __call__(
self, input_ids: npt.NDArray[np.intc], scores: npt.NDArray[np.single]
) -> npt.NDArray[np.single]:
for processor in self:
scores = processor(input_ids, scores)
return scores
StoppingCriteria = Callable[[npt.NDArray[np.intc], npt.NDArray[np.single]], bool]
class StoppingCriteriaList(List[StoppingCriteria]):
def __call__(
self, input_ids: npt.NDArray[np.intc], logits: npt.NDArray[np.single]
) -> bool:
return any([stopping_criteria(input_ids, logits) for stopping_criteria in self])