Add custom sampler order support (#5443)

This commit is contained in:
oobabooga 2024-02-06 11:20:10 -03:00 committed by GitHub
parent 7301c7618f
commit 8c35fefb3b
WARNING! Although there is a key with this ID in the database it does not verify this commit! This commit is SUSPICIOUS.
GPG key ID: B5690EEEBB952194
9 changed files with 205 additions and 113 deletions

View file

@ -55,8 +55,8 @@ For more information about the parameters, the [transformers documentation](http
* **mirostat_tau**: No idea, see the paper for details. According to the Preset Arena, 8 is a good value.
* **mirostat_eta**: No idea, see the paper for details. According to the Preset Arena, 0.1 is a good value.
* **dynamic_temperature**: Activates Dynamic Temperature. This modifies temperature to range between "dynatemp_low" (minimum) and "dynatemp_high" (maximum), with an entropy-based scaling. The steepness of the curve is controlled by "dynatemp_exponent".
* **smoothing_factor**: Activates Quadratic Sampling. This takes precedence over regular temperature and dynamic temperature, and replaces those samplers. When `0 < smoothing_factor < 1`, the logits distribution becomes flatter. When `smoothing_factor > 1`, it becomes more peaked.
* **temperature_last**: Makes temperature the last sampler instead of the first. With this, you can remove low probability tokens with a sampler like min_p and then use a high temperature to make the model creative without losing coherency.
* **smoothing_factor**: Activates Quadratic Sampling. When `0 < smoothing_factor < 1`, the logits distribution becomes flatter. When `smoothing_factor > 1`, it becomes more peaked.
* **temperature_last**: Makes temperature the last sampler instead of the first. With this, you can remove low probability tokens with a sampler like min_p and then use a high temperature to make the model creative without losing coherency. Note: this parameter takes precedence over "Sampler priority". That means that `temperature`/`dynamic_temperature`/`quadratic_sampling` will be removed from wherever they are and moved to the end of the stack.
* **do_sample**: When unchecked, sampling is entirely disabled, and greedy decoding is used instead (the most likely token is always picked).
* **Seed**: Set the Pytorch seed to this number. Note that some loaders do not use Pytorch (notably llama.cpp), and others are not deterministic (notably ExLlama v1 and v2). For these loaders, the seed has no effect.
* **encoder_repetition_penalty**: Also known as the "Hallucinations filter". Used to penalize tokens that are *not* in the prior text. Higher value = more likely to stay in context, lower value = more likely to diverge.
@ -77,6 +77,7 @@ To the right (or below if you are on mobile), the following parameters are prese
* **Add the bos_token to the beginning of prompts**: By default, the tokenizer will add a BOS (Beginning of Sequence) token to your prompt. During training, BOS tokens are used to separate different documents. If unchecked, no BOS token will be added, and the model will interpret your prompt as being in the middle of a document instead of at the start of one. This significantly changes the output and can make it more creative.
* **Skip special tokens**: When decoding the generated tokens, skip special tokens from being converted to their text representation. Otherwise, BOS appears as `<s>`, EOS as `</s>`, etc.
* **Activate text streaming**: When unchecked, the full response is outputted at once, without streaming the words one at a time. I recommend unchecking this parameter on high latency networks like running the webui on Google Colab or using `--share`.
* **Sampler priority**: Allows you to customize the order in which the different samplers are applied. The first sampler on the list gets applied first. With this, custom orders like `top_p -> temperature -> top_k` can be defined.
* **Load grammar from file**: Loads a GBNF grammar from a file under `text-generation-webui/grammars`. The output is written to the "Grammar" box below. You can also save and delete custom grammars using this menu.
* **Grammar**: Allows you to constrain the model output to a particular format. For instance, you can make the model generate lists, JSON, specific words, etc. Grammar is extremely powerful and I highly recommend it. The syntax looks a bit daunting at first sight, but it gets very easy once you understand it. See the [GBNF Guide](https://github.com/ggerganov/llama.cpp/blob/master/grammars/README.md) for details.

View file

@ -40,6 +40,7 @@ class GenerationOptions(BaseModel):
max_tokens_second: int = 0
prompt_lookup_num_tokens: int = 0
custom_token_bans: str = ""
sampler_priority: List[str] | str | None = Field(default=None, description="List of samplers where the first items will appear first in the stack. Example: [\"top_k\", \"temperature\", \"top_p\"].")
auto_max_new_tokens: bool = False
ban_eos_token: bool = False
add_bos_token: bool = True

View file

@ -182,6 +182,7 @@ def transformers_samplers():
'negative_prompt',
'ban_eos_token',
'custom_token_bans',
'sampler_priority',
'add_bos_token',
'skip_special_tokens',
'auto_max_new_tokens',
@ -230,6 +231,7 @@ loaders_samplers = {
'negative_prompt',
'ban_eos_token',
'custom_token_bans',
'sampler_priority',
'add_bos_token',
'skip_special_tokens',
'auto_max_new_tokens',
@ -287,6 +289,7 @@ loaders_samplers = {
'negative_prompt',
'ban_eos_token',
'custom_token_bans',
'sampler_priority',
'add_bos_token',
'skip_special_tokens',
'auto_max_new_tokens',

View file

@ -42,6 +42,7 @@ def default_preset():
'num_beams': 1,
'length_penalty': 1,
'early_stopping': False,
'sampler_priority': 'temperature\ndynamic_temperature\nquadratic_sampling\ntop_k\ntop_p\ntypical_p\nepsilon_cutoff\neta_cutoff\ntfs\ntop_a\nmin_p\nmirostat'
}

View file

@ -1,4 +1,5 @@
import math
import pprint
import torch
import transformers
@ -6,21 +7,21 @@ from transformers import LogitsWarper, is_torch_xpu_available
from transformers.generation.logits_process import (
LogitNormalization,
LogitsProcessor,
LogitsProcessorList,
TemperatureLogitsWarper
LogitsProcessorList
)
from modules import shared
from modules.logging_colors import logger
global_scores = None
class ModifiedTemperatureLogitsWarper(LogitsWarper):
class TemperatureLogitsWarperCustom(LogitsWarper):
'''
Based on the original Transformers temperature logits warper, this
adds support for dynamic temperature and quadratic sampling.
A copy of the original Transformers temperature logits warper.
'''
def __init__(self, temperature: float, dynamic_temperature: bool, dynatemp_low: float, dynatemp_high: float, dynatemp_exponent: float, smoothing_factor: float):
def __init__(self, temperature: float):
if not isinstance(temperature, float) or not (temperature > 0):
except_msg = (
f"`temperature` (={temperature}) has to be a strictly positive float, otherwise your next token "
@ -32,81 +33,90 @@ class ModifiedTemperatureLogitsWarper(LogitsWarper):
raise ValueError(except_msg)
self.temperature = temperature
self.dynamic_temperature = dynamic_temperature
def __call__(self, input_ids: torch.LongTensor, scores: torch.FloatTensor) -> torch.FloatTensor:
scores = scores / self.temperature
return scores
class DynamicTemperatureLogitsWarper(LogitsWarper):
'''
Dynamic temperature.
'''
def __init__(self, dynatemp_low: float, dynatemp_high: float, dynatemp_exponent: float):
self.dynatemp_low = dynatemp_low
self.dynatemp_high = dynatemp_high
self.dynatemp_exponent = dynatemp_exponent
def __call__(self, input_ids: torch.LongTensor, scores: torch.FloatTensor) -> torch.FloatTensor:
min_temp = self.dynatemp_low
max_temp = self.dynatemp_high
exponent_val = self.dynatemp_exponent
# Convert logits to probabilities
probs = torch.softmax(scores, dim=-1)
# Calculate entropy of the softmax probabilities
entropy = -1.0 * torch.where(probs > 0, probs * torch.log(probs), torch.zeros_like(probs)).sum()
# Guard against future possible division by zero
entropy = max(entropy, torch.tensor(1e-10)) # Ensures entropy is slightly greater than 0
# Any logits which are not -Infinity will be considered for calculating max entropy.
num_valid_tokens = torch.sum(scores > -float('inf')).item()
# Now, calculate the max entropy by using only the valid tokens' count
max_entropy = math.log(num_valid_tokens)
# Guard against future possible division by zero
max_entropy = max_entropy if max_entropy > 0.0 else 1e-10
# Normalize the entropy
normalized_entropy = entropy / max_entropy
# Map the normalized entropy to the desired temperature range using the power function
dyn_temp = min_temp + (max_temp - min_temp) * (normalized_entropy.pow(exponent_val))
# Apply the dynamically calculated temperature scaling
scores = scores / dyn_temp
# print("----------------------\nTemperature from generation_config:", self.temperature)
# print("min_temp:", min_temp)
# print("max_temp:", max_temp)
# print("Entropy:", entropy.item())
# print("Max Possible Entropy considering valid tokens only:", max_entropy)
# print("Normalized Entropy:", normalized_entropy.item())
# print("Dynamic Temperature (dyn_temp):", dyn_temp.item())
# print("----------------------")
# max_prob_token_id = torch.argmax(scores, dim=-1) # Get the token ID with the highest probability
# max_prob_token = shared.tokenizer.convert_ids_to_tokens(int(max_prob_token_id)) # Convert ID to token
# print("--- T=", float(dyn_temp), "token=", max_prob_token, "min=", min_temp, "max=", max_temp, "exponent=", exponent_val)
return scores
class QuadraticSamplingLogitsWarper(LogitsWarper):
'''
Quadratic sampling.
'''
def __init__(self, smoothing_factor: float):
self.smoothing_factor = smoothing_factor
def __call__(self, input_ids: torch.LongTensor, scores: torch.FloatTensor) -> torch.FloatTensor:
# Compute the maximum logit value
max_logit = scores.max()
# Quadratic sampling
if self.smoothing_factor > 0:
# Apply the quadratic transformation
transformed_logits = -(self.smoothing_factor * (scores - max_logit)**2) + max_logit
# Compute the maximum logit value
max_logit = scores.max()
# No need to print the top 5 logits since this is not required
# print("Original top 5 logits: ", torch.topk(scores, 5))
# print("New top 5 logits: ", torch.topk(transformed_logits, 5))
# Apply the quadratic transformation
transformed_logits = -(self.smoothing_factor * (scores - max_logit)**2) + max_logit
# No need to print the top 5 logits since this is not required
# print("Original top 5 logits: ", torch.topk(scores, 5))
# print("New top 5 logits: ", torch.topk(transformed_logits, 5))
return transformed_logits
# Dynamic temperature
elif self.dynamic_temperature:
min_temp = self.dynatemp_low
max_temp = self.dynatemp_high
exponent_val = self.dynatemp_exponent
# Convert logits to probabilities
probs = torch.softmax(scores, dim=-1)
# Calculate entropy of the softmax probabilities
entropy = -1.0 * torch.where(probs > 0, probs * torch.log(probs), torch.zeros_like(probs)).sum()
# Guard against future possible division by zero
entropy = max(entropy, torch.tensor(1e-10)) # Ensures entropy is slightly greater than 0
# Any logits which are not -Infinity will be considered for calculating max entropy.
num_valid_tokens = torch.sum(scores > -float('inf')).item()
# Now, calculate the max entropy by using only the valid tokens' count
max_entropy = math.log(num_valid_tokens)
# Guard against future possible division by zero
max_entropy = max_entropy if max_entropy > 0.0 else 1e-10
# Normalize the entropy
normalized_entropy = entropy / max_entropy
# Map the normalized entropy to the desired temperature range using the power function
dyn_temp = min_temp + (max_temp - min_temp) * (normalized_entropy.pow(exponent_val))
# Apply the dynamically calculated temperature scaling
scores = scores / dyn_temp
# print("----------------------\nTemperature from generation_config:", self.temperature)
# print("min_temp:", min_temp)
# print("max_temp:", max_temp)
# print("Entropy:", entropy.item())
# print("Max Possible Entropy considering valid tokens only:", max_entropy)
# print("Normalized Entropy:", normalized_entropy.item())
# print("Dynamic Temperature (dyn_temp):", dyn_temp.item())
# print("----------------------")
# max_prob_token_id = torch.argmax(scores, dim=-1) # Get the token ID with the highest probability
# max_prob_token = shared.tokenizer.convert_ids_to_tokens(int(max_prob_token_id)) # Convert ID to token
# print("--- T=", float(dyn_temp), "token=", max_prob_token, "min=", min_temp, "max=", max_temp, "exponent=", exponent_val)
return scores
# Regular temperature
else:
scores = scores / self.temperature
return scores
return transformed_logits
class MinPLogitsWarper(LogitsWarper):
@ -209,6 +219,7 @@ class MirostatLogitsWarper(LogitsWarper):
def __init__(self, mirostat_mode: int, mirostat_tau: float, mirostat_eta: float, filter_value: float = -float("Inf"), min_tokens_to_keep: int = 1):
if mirostat_mode not in [2]:
raise ValueError(f"`mirostat` has to be a an integer 2, but is {mirostat_mode}")
self.mirostat_mode = mirostat_mode
self.mirostat_eta = mirostat_eta
self.mirostat_tau = mirostat_tau
@ -301,44 +312,74 @@ class RepetitionPenaltyLogitsProcessorWithRange(LogitsProcessor):
def get_logits_warper_patch(self, generation_config):
# Make sure that temperature is float and not int
# Parameter sanitization
if isinstance(generation_config.temperature, int):
generation_config.temperature = float(generation_config.temperature)
temperature = generation_config.temperature
if generation_config.dynamic_temperature or generation_config.smoothing_factor > 0:
# Make sure TemperatureLogitsWarper will be created by temporarily
# setting temperature to a value != 1.
generation_config.temperature = 1.1
generation_config.temperature = float(generation_config.temperature) # Must be float
# Get the original warpers
warpers = self._get_logits_warper_old(generation_config)
# Replace temperature with our modified class.
# Currently, it behaves identically to the original.
for i in range(len(warpers)):
if warpers[i].__class__.__name__ == 'TemperatureLogitsWarper':
warpers[i] = ModifiedTemperatureLogitsWarper(
temperature,
generation_config.dynamic_temperature,
generation_config.dynatemp_low,
generation_config.dynatemp_high,
generation_config.dynatemp_exponent,
generation_config.smoothing_factor
warpers[i] = TemperatureLogitsWarperCustom(
generation_config.temperature,
)
# Add custom warpers
warpers_to_add = LogitsProcessorList()
min_tokens_to_keep = 2 if generation_config.num_beams > 1 else 1
if generation_config.tfs is not None and 0.0 <= generation_config.tfs < 1.0:
warpers_to_add.append(
TailFreeLogitsWarper(
tfs=generation_config.tfs,
min_tokens_to_keep=min_tokens_to_keep
)
)
if generation_config.top_a is not None and 0.0 < generation_config.top_a <= 1.0:
warpers_to_add.append(
TopALogitsWarper(
top_a=generation_config.top_a,
min_tokens_to_keep=min_tokens_to_keep
)
)
if generation_config.min_p is not None and 0.0 < generation_config.min_p <= 1.0:
warpers_to_add.append(
MinPLogitsWarper(
min_p=generation_config.min_p,
min_tokens_to_keep=min_tokens_to_keep
)
)
if generation_config.dynamic_temperature:
warpers_to_add.append(
DynamicTemperatureLogitsWarper(
dynatemp_low=generation_config.dynatemp_low,
dynatemp_high=generation_config.dynatemp_high,
dynatemp_exponent=generation_config.dynatemp_exponent,
)
)
if generation_config.smoothing_factor > 0:
warpers_to_add.append(
QuadraticSamplingLogitsWarper(
smoothing_factor=generation_config.smoothing_factor
)
)
if generation_config.mirostat_mode is not None and generation_config.mirostat_mode == 2:
warpers_to_add.append(MirostatLogitsWarper(mirostat_mode=generation_config.mirostat_mode, mirostat_eta=generation_config.mirostat_eta, mirostat_tau=generation_config.mirostat_tau, min_tokens_to_keep=min_tokens_to_keep))
# We need to disable samplers other than temperature
for warper in warpers:
if not isinstance(warper, TemperatureLogitsWarper):
warpers.remove(warper)
else:
if generation_config.tfs is not None and 0.0 <= generation_config.tfs < 1.0:
warpers_to_add.append(TailFreeLogitsWarper(tfs=generation_config.tfs, min_tokens_to_keep=min_tokens_to_keep))
if generation_config.top_a is not None and 0.0 < generation_config.top_a <= 1.0:
warpers_to_add.append(TopALogitsWarper(top_a=generation_config.top_a, min_tokens_to_keep=min_tokens_to_keep))
if generation_config.min_p is not None and 0.0 < generation_config.min_p <= 1.0:
warpers_to_add.append(MinPLogitsWarper(min_p=generation_config.min_p, min_tokens_to_keep=min_tokens_to_keep))
warpers_to_add.append(
MirostatLogitsWarper(
mirostat_mode=generation_config.mirostat_mode,
mirostat_eta=generation_config.mirostat_eta,
mirostat_tau=generation_config.mirostat_tau,
min_tokens_to_keep=min_tokens_to_keep
)
)
if len(warpers) > 0 and isinstance(warpers[-1], LogitNormalization):
normalize = warpers.pop(-1)
@ -346,23 +387,57 @@ def get_logits_warper_patch(self, generation_config):
normalize = None
warpers += warpers_to_add
if generation_config.temperature_last:
temperature_idx = None
for i in range(len(warpers)):
if warpers[i].__class__.__name__ in ['TemperatureLogitsWarper', 'ModifiedTemperatureLogitsWarper']:
temperature_idx = i
break
if temperature_idx is not None:
warpers.append(warpers.pop(temperature_idx))
# Sort the samplers.
sampler_priority = generation_config.sampler_priority
# Handle temperature_last
if generation_config.temperature_last:
for param_name in ['temperature', 'dynamic_temperature', 'quadratic_sampling']:
if param_name in sampler_priority:
if param_name in sampler_priority:
index = sampler_priority.index(param_name)
sampler_priority.append(sampler_priority.pop(index))
else:
sampler_priority.append(param_name)
class_name_to_nickname = {
'DynamicTemperatureLogitsWarper': 'dynamic_temperature',
'EpsilonLogitsWarper': 'epsilon_cutoff',
'EtaLogitsWarper': 'eta_cutoff',
'MinPLogitsWarper': 'min_p',
'MirostatLogitsWarper': 'mirostat',
'QuadraticSamplingLogitsWarper': 'quadratic_sampling',
'TailFreeLogitsWarper': 'tfs',
'TemperatureLogitsWarperCustom': 'temperature',
'TopALogitsWarper': 'top_a',
'TopKLogitsWarper': 'top_k',
'TopPLogitsWarper': 'top_p',
'TypicalLogitsWarper': 'typical_p'
}
def custom_sort_key(obj):
class_name = obj.__class__.__name__
# Return a large value if class name is not mapped or if the mapped nickname is not in priority
if class_name not in class_name_to_nickname or class_name_to_nickname[class_name] not in sampler_priority:
return float('inf')
# Return the index of the nickname in the priority list for sorting
return sampler_priority.index(class_name_to_nickname[class_name])
# Sort the list using the custom key function
warpers = sorted(warpers, key=custom_sort_key)
if normalize is not None:
warpers.append(normalize)
warpers.append(SpyLogitsWarper())
warpers = LogitsProcessorList(warpers)
# for i in range(len(warpers)):
# print(warpers[i].__class__.__name__)
if shared.args.verbose:
logger.info("WARPERS=")
pprint.PrettyPrinter(indent=4, sort_dicts=False).pprint([x.__class__.__name__ for x in warpers])
return warpers
@ -402,6 +477,7 @@ def generation_config_init_patch(self, **kwargs):
self.presence_penalty = kwargs.pop("presence_penalty", 0)
self.frequency_penalty = kwargs.pop("frequency_penalty", 0)
self.temperature_last = kwargs.pop("temperature_last", False)
self.sampler_priority = kwargs.pop("sampler_priority", ['temperature', 'dynamic_temperature', 'quadratic_sampling', 'top_k', 'top_p', 'typical_p', 'epsilon_cutoff', 'eta_cutoff', 'tfs', 'top_a', 'min_p', 'mirostat'])
def hijack_samplers():

View file

@ -50,6 +50,7 @@ settings = {
'prompt_lookup_num_tokens': 0,
'custom_stopping_strings': '',
'custom_token_bans': '',
'sampler_priority': 'temperature,top_k,top_p,typical_p,epsilon_cutoff,eta_cutoff,tfs,top_a,min_p,dynamic_temperature,quadratic_sampling,mirostat',
'auto_max_new_tokens': False,
'ban_eos_token': False,
'add_bos_token': True,

View file

@ -291,6 +291,11 @@ def generate_reply_HF(question, original_question, seed, state, stopping_strings
if k in state:
generate_params[k] = state[k]
if isinstance(state['sampler_priority'], list):
generate_params['sampler_priority'] = state['sampler_priority']
elif isinstance(state['sampler_priority'], str):
generate_params['sampler_priority'] = [x.strip() for x in state['sampler_priority'].replace('\n', ',').split(',') if x.strip()]
if state['negative_prompt'] != '':
generate_params['negative_prompt_ids'] = encode(state['negative_prompt'])

View file

@ -149,6 +149,7 @@ def list_interface_input_elements():
'add_bos_token',
'ban_eos_token',
'custom_token_bans',
'sampler_priority',
'truncation_length',
'custom_stopping_strings',
'skip_special_tokens',

View file

@ -49,12 +49,12 @@ def create_ui(default_preset):
shared.gradio['mirostat_mode'] = gr.Slider(0, 2, step=1, value=generate_params['mirostat_mode'], label='mirostat_mode', info='mode=1 is for llama.cpp only.')
shared.gradio['mirostat_tau'] = gr.Slider(0, 10, step=0.01, value=generate_params['mirostat_tau'], label='mirostat_tau')
shared.gradio['mirostat_eta'] = gr.Slider(0, 1, step=0.01, value=generate_params['mirostat_eta'], label='mirostat_eta')
shared.gradio['smoothing_factor'] = gr.Slider(0.0, 10.0, value=generate_params['smoothing_factor'], step=0.01, label='smoothing_factor', info='Replaces temperature with Quadratic Sampling.')
shared.gradio['smoothing_factor'] = gr.Slider(0.0, 10.0, value=generate_params['smoothing_factor'], step=0.01, label='smoothing_factor', info='Activates Quadratic Sampling.')
shared.gradio['dynamic_temperature'] = gr.Checkbox(value=generate_params['dynamic_temperature'], label='dynamic_temperature')
shared.gradio['dynatemp_low'] = gr.Slider(0.01, 5, value=generate_params['dynatemp_low'], step=0.01, label='dynatemp_low', visible=generate_params['dynamic_temperature'])
shared.gradio['dynatemp_high'] = gr.Slider(0.01, 5, value=generate_params['dynatemp_high'], step=0.01, label='dynatemp_high', visible=generate_params['dynamic_temperature'])
shared.gradio['dynatemp_exponent'] = gr.Slider(0.01, 5, value=generate_params['dynatemp_exponent'], step=0.01, label='dynatemp_exponent', visible=generate_params['dynamic_temperature'])
shared.gradio['temperature_last'] = gr.Checkbox(value=generate_params['temperature_last'], label='temperature_last', info='Makes temperature the last sampler instead of the first.')
shared.gradio['temperature_last'] = gr.Checkbox(value=generate_params['temperature_last'], label='temperature_last', info='Moves temperature/dynamic temperature/quadratic sampling to the end of the sampler stack, ignoring their positions in "Sampler priority".')
shared.gradio['do_sample'] = gr.Checkbox(value=generate_params['do_sample'], label='do_sample')
shared.gradio['seed'] = gr.Number(value=shared.settings['seed'], label='Seed (-1 for random)')
with gr.Accordion('Other parameters', open=False):
@ -85,6 +85,9 @@ def create_ui(default_preset):
shared.gradio['skip_special_tokens'] = gr.Checkbox(value=shared.settings['skip_special_tokens'], label='Skip special tokens', info='Some specific models need this unset.')
shared.gradio['stream'] = gr.Checkbox(value=shared.settings['stream'], label='Activate text streaming')
with gr.Blocks():
shared.gradio['sampler_priority'] = gr.Textbox(value=generate_params['sampler_priority'], lines=12, label='Sampler priority', info='Parameter names separated by new lines or commas.')
with gr.Row() as shared.gradio['grammar_file_row']:
shared.gradio['grammar_file'] = gr.Dropdown(value='None', choices=utils.get_available_grammars(), label='Load grammar from file (.gbnf)', elem_classes='slim-dropdown')
ui.create_refresh_button(shared.gradio['grammar_file'], lambda: None, lambda: {'choices': utils.get_available_grammars()}, 'refresh-button', interactive=not mu)