Transformers documentation

Youtu-LLM

Transformers

You are viewing main version, which requires installation from source. If you'd like regular pip install, checkout the latest stable version (v5.0.0).

Join the Hugging Face community

and get access to the augmented documentation experience

Collaborate on models, datasets and Spaces

Faster examples with accelerated inference

Switch between documentation themes

to get started

This model was released on 2025-12-31 and added to Hugging Face Transformers on 2026-01-28.

Youtu-LLM

Overview

The Youtu-LLM model was proposed in Youtu-LLM Technical Report by Tencent Youtu Team.

The abstract from the paper is the following: We introduce Youtu-LLM, a lightweight yet powerful language model that harmonizes high computational efficiency with native agentic intelligence. Unlike typical small models that rely on distillation, Youtu-LLM (1.96B) is pre-trained from scratch to systematically cultivate reasoning and planning capabilities. The key technical advancements are as follows: (1) Compact Architecture with Long-Context Support: Built on a dense Multi-Latent Attention (MLA) architecture with a novel STEM-oriented vocabulary, Youtu-LLM supports a 128k context window. This design enables robust long-context reasoning and state tracking within a minimal memory footprint, making it ideal for long-horizon agent and reasoning tasks. (2) Principled “Commonsense-STEM-Agent” Curriculum: We curated a massive corpus of approximately 11T tokens and implemented a multi-stage training strategy. By progressively shifting the pre-training data distribution from general commonsense to complex STEM and agentic tasks, we ensure the model acquires deep cognitive abilities rather than superficial alignment. (3) Scalable Agentic Mid-training: Specifically for the agentic mid-training, we employ diverse data construction schemes to synthesize rich and varied trajectories across math, coding, and tool-use domains. This high-quality data enables the model to internalize planning and reflection behaviors effectively. Extensive evaluations show that Youtu-LLM sets a new state-of-the-art for sub-2B LLMs. On general benchmarks, it achieves competitive performance against larger models, while on agent-specific tasks, it significantly surpasses existing SOTA baselines, demonstrating that lightweight models can possess strong intrinsic agentic capabilities.

Usage tips

The model uses Multi-head Latent Attention (MLA) architectures for efficient inference. The model can be used for various language tasks after being pre-trained on approximate 11 trillion tokens and going through Supervised Fine-Tuning and Reinforcement Learning stages. The following example demonstrates how to load the model, enable Reasoning Mode, and use the re module to parse the “Thought Process” and the “Final Answer” from the output.

import re
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM

# 1. Configure Model
model_id = "tencent/Youtu-LLM-2B"

# 2. Initialize Tokenizer and Model
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    device_map="auto"
)

# 3. Construct Dialogue Input
prompt = "Hello"
messages = [{"role": "user", "content": prompt}]

# Use apply_chat_template to construct input; set enable_thinking=True to activate Reasoning Mode
input_ids = tokenizer.apply_chat_template(
    messages, 
    tokenize=True, 
    add_generation_prompt=True, 
    return_tensors="pt",
    enable_thinking=True
).to(model.device)

# 4. Generate Response
outputs = model.generate(
    input_ids,
    max_new_tokens=512,
    do_sample=True,
    temperature=1.0,
    top_p=0.95,
    repetition_penalty=1.05
)

# 5. Parse Results
full_response = tokenizer.decode(outputs[0], skip_special_tokens=True)

def parse_reasoning(text):
    """Extract thought process within <think> tags and the subsequent answer content"""
    thought_pattern = r"<think>(.*?)</think>"
    match = re.search(thought_pattern, text, re.DOTALL)
    
    if match:
        thought = match.group(1).strip()
        answer = text.split("</think>")[-1].strip()
    else:
        thought = "(No explicit thought process generated)"
        answer = text
    return thought, answer

thought, final_answer = parse_reasoning(full_response)

print(f"\n{'='*20} Thought Process {'='*20}\n{thought}")
print(f"\n{'='*20} Final Answer {'='*20}\n{final_answer}")

This generated:

==================== Thought Process ====================
The user greeted with 'Hello', which is a simple and friendly opening. Since the input is in English, I should respond in English as per the instruction. I need to introduce myself clearly according to the defined identity: state my name (Youtu-llm), developer (Tencent Youtu team), purpose (helping users solve problems), key capabilities (mathematics, coding, Agent), and goal (efficient and accurate problem-solving). The response should be welcoming and open-ended to encourage further interaction, while staying within the provided identity constraints. No extra information beyond what is specified should be added.

==================== Final Answer ====================
Hello! I am Youtu-llm, a large language model developed by the Tencent Youtu team. I am designed to assist users in solving various problems, excelling in tasks such as mathematics, coding, and Agent-related operations. My goal is to make problem-solving more efficient and accurate through intelligent interaction. How can I assist you today?

Key Configuration Details

Reasoning Mode Toggle

Controlled via the enable_thinking parameter in the apply_chat_template method:

True (Recommended Default): Activates Chain of Thought; ideal for complex logic and reasoning tasks.
False: Outputs results directly; faster response time, suitable for simple conversations.

Recommended Decoding Parameters

Depending on your use case, we suggest adjusting the following hyperparameters for optimal generation:

Parameter	Reasoning Mode	Normal Mode
`do_sample`	`True`	`True`
`temperature`	1.0 (Maintains creativity)	0.7 (More stable results)
`top_p`	0.95	0.8
`top_k`	20	20
`repetition_penalty`	1.05	-

Tip: When using Reasoning Mode, a higher temperature helps the model perform deeper, more divergent thinking.

YoutuConfig

class transformers.YoutuConfig

< source >

Parameters

vocab_size (int, optional, defaults to 128256) — Vocabulary size of the Deep model. Defines the number of different tokens that can be represented by the inputs_ids passed when calling YoutuModel
hidden_size (int, optional, defaults to 2048) — Dimension of the hidden representations.
intermediate_size (int, optional, defaults to 6144) — Dimension of the MLP representations.
num_hidden_layers (int, optional, defaults to 32) — Number of hidden layers in the Transformer decoder.
num_attention_heads (int, optional, defaults to 16) — Number of attention heads for each attention layer in the Transformer decoder.
num_key_value_heads (int, optional, defaults to 16) — In MLA, num_key_value_heads=num_attention_heads.
kv_lora_rank (int, optional, defaults to 512) — Rank of the LoRA matrices for key and value projections.
q_lora_rank (int, optional, defaults to 1536) — Rank of the LoRA matrices for query projections.
qk_rope_head_dim (int, optional, defaults to 64) — Dimension of the query/key heads that use rotary position embeddings.
v_head_dim (int, optional, defaults to 128) — Dimension of the value heads.
qk_nope_head_dim (int, optional, defaults to 128) — Dimension of the query/key heads that don’t use rotary position embeddings.
hidden_act (str or function, optional, defaults to "silu") — The non-linear activation function (function or string) in the decoder.
max_position_embeddings (int, optional, defaults to 131072) — The maximum sequence length that this model might ever be used with.
initializer_range (float, optional) — The standard deviation of the truncated_normal_initializer for initializing all weight matrices, except embedding matrices.
embedding_initializer_range (float, optional) — The standard deviation of the truncated_normal_initializer for initializing all embedding matrices.
rms_norm_eps (float, optional, defaults to 1e-06) — The epsilon used by the rms normalization layers.
use_cache (bool, optional, defaults to True) — Whether or not the model should return the last key/values attentions (not used by all models). Only relevant if config.is_decoder=True.
pad_token_id (int, optional) — Padding token id.
bos_token_id (int, optional, defaults to 128000) — Beginning of stream token id.
eos_token_id (int, optional, defaults to 128001) — End of stream token id.
tie_word_embeddings (bool, optional, defaults to True) — Whether to tie weight embeddings
rope_parameters (RopeParameters, optional) — Dictionary containing the configuration parameters for the RoPE embeddings. The dictionary should contain a value for rope_theta and optionally parameters used for scaling in case you want to use RoPE with longer max_position_embeddings.
rope_interleave (bool, optional, defaults to True) — Whether to interleave the rotary position embeddings.
attention_bias (bool, defaults to False, optional, defaults to False) — Whether to use a bias in the query, key, value and output projection layers during self-attention.
attention_dropout (float, optional, defaults to 0.0) — The dropout ratio for the attention probabilities.

This is the configuration class to store the configuration of a YoutuModel. It is used to instantiate an Youtu model according to the specified arguments, defining the model architecture. Instantiating a configuration with the defaults will yield a similar configuration to that of the Youtu-LLM-2B. e.g. tencent/Youtu-LLM-2B

Configuration objects inherit from PreTrainedConfig and can be used to control the model outputs. Read the documentation from PreTrainedConfig for more information.

>>> from transformers import YoutuModel, YoutuConfig
>>> # Initializing a Youtu-LLM-2B style configuration
>>> configuration = YoutuConfig()
>>> # Accessing the model configuration
>>> configuration = model.config

YoutuModel

class transformers.YoutuModel

< source >

( config: YoutuConfig )

Parameters

config (YoutuConfig) — Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out the from_pretrained() method to load the model weights.

The bare Youtu Model outputting raw hidden-states without any specific head on top.

This model inherits from PreTrainedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.)

This model is also a PyTorch torch.nn.Module subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.

forward

< source >

( input_ids: torch.LongTensor | None = None attention_mask: torch.Tensor | None = None position_ids: torch.LongTensor | None = None past_key_values: transformers.cache_utils.Cache | None = None inputs_embeds: torch.FloatTensor | None = None cache_position: torch.LongTensor | None = None use_cache: bool | None = None **kwargs: typing_extensions.Unpack[transformers.utils.generic.TransformersKwargs] ) → transformers.modeling_outputs.BaseModelOutputWithPast or tuple(torch.FloatTensor)

Parameters

input_ids (torch.LongTensor of shape (batch_size, sequence_length), optional) — Indices of input sequence tokens in the vocabulary. Padding will be ignored by default.

Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details.

What are input IDs?
attention_mask (torch.Tensor of shape (batch_size, sequence_length), optional) — Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:
- 1 for tokens that are not masked,
- 0 for tokens that are masked.
What are attention masks?
position_ids (torch.LongTensor of shape (batch_size, sequence_length), optional) — Indices of positions of each input sequence tokens in the position embeddings. Selected in the range [0, config.n_positions - 1].

What are position IDs?
past_key_values (~cache_utils.Cache, optional) — Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used to speed up sequential decoding. This typically consists in the past_key_values returned by the model at a previous stage of decoding, when use_cache=True or config.use_cache=True.

Only Cache instance is allowed as input, see our kv cache guide. If no past_key_values are passed, DynamicCache will be initialized by default.

The model will output the same cache format that is fed as input.

If past_key_values are used, the user is expected to input only unprocessed input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, unprocessed_length) instead of all input_ids of shape (batch_size, sequence_length).
inputs_embeds (torch.FloatTensor of shape (batch_size, sequence_length, hidden_size), optional) — Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.
cache_position (torch.LongTensor of shape (sequence_length), optional) — Indices depicting the position of the input sequence tokens in the sequence. Contrarily to position_ids, this tensor is not affected by padding. It is used to update the cache in the correct position and to infer the complete sequence length.
use_cache (bool, optional) — If set to True, past_key_values key value states are returned and can be used to speed up decoding (see past_key_values).

Returns

transformers.modeling_outputs.BaseModelOutputWithPast or tuple(torch.FloatTensor)

A transformers.modeling_outputs.BaseModelOutputWithPast or a tuple of torch.FloatTensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (YoutuConfig) and inputs.

last_hidden_state (torch.FloatTensor of shape (batch_size, sequence_length, hidden_size)) — Sequence of hidden-states at the output of the last layer of the model.

If past_key_values is used only the last hidden-state of the sequences of shape (batch_size, 1, hidden_size) is output.
past_key_values (Cache, optional, returned when use_cache=True is passed or when config.use_cache=True) — It is a Cache instance. For more details, see our kv cache guide.

Contains pre-computed hidden-states (key and values in the self-attention blocks and optionally if config.is_encoder_decoder=True in the cross-attention blocks) that can be used (see past_key_values input) to speed up sequential decoding.
hidden_states (tuple(torch.FloatTensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) — Tuple of torch.FloatTensor (one for the output of the embeddings, if the model has an embedding layer, + one for the output of each layer) of shape (batch_size, sequence_length, hidden_size).

Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
attentions (tuple(torch.FloatTensor), optional, returned when output_attentions=True is passed or when config.output_attentions=True) — Tuple of torch.FloatTensor (one for each layer) of shape (batch_size, num_heads, sequence_length, sequence_length).

Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.

The YoutuModel forward method, overrides the __call__ special method.

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.

YoutuForCausalLM

class transformers.YoutuForCausalLM

< source >

( config )

Parameters

config (YoutuForCausalLM) — Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out the from_pretrained() method to load the model weights.

The Youtu Model for causal language modeling.

This model is also a PyTorch torch.nn.Module subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.

forward

< source >

( input_ids: torch.LongTensor | None = None attention_mask: torch.Tensor | None = None position_ids: torch.LongTensor | None = None past_key_values: transformers.cache_utils.Cache | None = None inputs_embeds: torch.FloatTensor | None = None labels: torch.LongTensor | None = None use_cache: bool | None = None cache_position: torch.LongTensor | None = None logits_to_keep: int | torch.Tensor = 0 **kwargs: typing_extensions.Unpack[transformers.utils.generic.TransformersKwargs] ) → transformers.modeling_outputs.CausalLMOutputWithPast or tuple(torch.FloatTensor)

Parameters

input_ids (torch.LongTensor of shape (batch_size, sequence_length), optional) — Indices of input sequence tokens in the vocabulary. Padding will be ignored by default.

Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details.

What are input IDs?
attention_mask (torch.Tensor of shape (batch_size, sequence_length), optional) — Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:
- 1 for tokens that are not masked,
- 0 for tokens that are masked.
What are attention masks?
position_ids (torch.LongTensor of shape (batch_size, sequence_length), optional) — Indices of positions of each input sequence tokens in the position embeddings. Selected in the range [0, config.n_positions - 1].

What are position IDs?
past_key_values (~cache_utils.Cache, optional) — Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used to speed up sequential decoding. This typically consists in the past_key_values returned by the model at a previous stage of decoding, when use_cache=True or config.use_cache=True.

Only Cache instance is allowed as input, see our kv cache guide. If no past_key_values are passed, DynamicCache will be initialized by default.

The model will output the same cache format that is fed as input.

If past_key_values are used, the user is expected to input only unprocessed input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, unprocessed_length) instead of all input_ids of shape (batch_size, sequence_length).
inputs_embeds (torch.FloatTensor of shape (batch_size, sequence_length, hidden_size), optional) — Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.
labels (torch.LongTensor of shape (batch_size, sequence_length), optional) — Labels for computing the masked language modeling loss. Indices should either be in [0, ..., config.vocab_size] or -100 (see input_ids docstring). Tokens with indices set to -100 are ignored (masked), the loss is only computed for the tokens with labels in [0, ..., config.vocab_size].
use_cache (bool, optional) — If set to True, past_key_values key value states are returned and can be used to speed up decoding (see past_key_values).
cache_position (torch.LongTensor of shape (sequence_length), optional) — Indices depicting the position of the input sequence tokens in the sequence. Contrarily to position_ids, this tensor is not affected by padding. It is used to update the cache in the correct position and to infer the complete sequence length.
logits_to_keep (Union[int, torch.Tensor], optional, defaults to 0) — If an int, compute logits for the last logits_to_keep tokens. If 0, calculate logits for all input_ids (special case). Only last token logits are needed for generation, and calculating them only for that token can save memory, which becomes pretty significant for long sequences or large vocabulary size. If a torch.Tensor, must be 1D corresponding to the indices to keep in the sequence length dimension. This is useful when using packed tensor format (single dimension for batch and sequence length).

Returns

transformers.modeling_outputs.CausalLMOutputWithPast or tuple(torch.FloatTensor)

A transformers.modeling_outputs.CausalLMOutputWithPast or a tuple of torch.FloatTensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (YoutuConfig) and inputs.

loss (torch.FloatTensor of shape (1,), optional, returned when labels is provided) — Language modeling loss (for next-token prediction).
logits (torch.FloatTensor of shape (batch_size, sequence_length, config.vocab_size)) — Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).
past_key_values (Cache, optional, returned when use_cache=True is passed or when config.use_cache=True) — It is a Cache instance. For more details, see our kv cache guide.

Contains pre-computed hidden-states (key and values in the self-attention blocks) that can be used (see past_key_values input) to speed up sequential decoding.
hidden_states (tuple(torch.FloatTensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) — Tuple of torch.FloatTensor (one for the output of the embeddings, if the model has an embedding layer, + one for the output of each layer) of shape (batch_size, sequence_length, hidden_size).

Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
attentions (tuple(torch.FloatTensor), optional, returned when output_attentions=True is passed or when config.output_attentions=True) — Tuple of torch.FloatTensor (one for each layer) of shape (batch_size, num_heads, sequence_length, sequence_length).

Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.

The YoutuForCausalLM forward method, overrides the __call__ special method.

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.

Example:

>>> from transformers import AutoTokenizer, YoutuForCausalLM

>>> model = YoutuForCausalLM.from_pretrained("meta-youtu/Youtu-2-7b-hf")
>>> tokenizer = AutoTokenizer.from_pretrained("meta-youtu/Youtu-2-7b-hf")

>>> prompt = "Hey, are you conscious? Can you talk to me?"
>>> inputs = tokenizer(prompt, return_tensors="pt")

>>> # Generate
>>> generate_ids = model.generate(inputs.input_ids, max_length=30)
>>> tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
"Hey, are you conscious? Can you talk to me?\nI'm not conscious, but I can talk to you."

Update on GitHub

←YOSO Zamba→