# coding=utf-8
from functools import partial
from typing import Callable, List, Optional, Tuple, Union

import torch
import torch.nn.functional as F
from torch import nn

from transformers.cache_utils import Cache, DynamicCache, SlidingWindowCache, StaticCache
from transformers.generation import GenerationMixin
from transformers.modeling_attn_mask_utils import AttentionMaskConverter
from transformers.modeling_flash_attention_utils import FlashAttentionKwargs
from transformers.modeling_outputs import (
    MoeCausalLMOutputWithPast,
    MoeModelOutputWithPast
)
from transformers.modeling_rope_utils import ROPE_INIT_FUNCTIONS, dynamic_rope_update
from transformers.modeling_utils import ALL_ATTENTION_FUNCTIONS, PreTrainedModel
from transformers.processing_utils import Unpack
from transformers.utils import LossKwargs, can_return_tuple, is_torch_flex_attn_available, logging
from .configuration_smallthinker import SmallthinkerConfig


if is_torch_flex_attn_available():
    from torch.nn.attention.flex_attention import BlockMask

    from transformers.integrations.flex_attention import make_flex_block_causal_mask


logger = logging.get_logger(__name__)


class SmallthinkerHierarchicalMLP(nn.Module):
    def __init__(self, config: SmallthinkerConfig):
        super().__init__()
        self.config = config
        self.hidden_dim = config.hidden_size
        self.ffn_dim = config.moe_ffn_hidden_size
        self.moe_enable_secondary_experts = config.moe_enable_secondary_experts
        if self.moe_enable_secondary_experts:
            self.num_secondary_experts = config.moe_num_secondary_experts
            self.secondary_expert_size = config.moe_secondary_expert_size
            self.secondary_gate = nn.Linear(self.hidden_dim, self.num_secondary_experts, bias=False)

        self.up = nn.Linear(self.hidden_dim, self.ffn_dim, bias=False)
        self.gate = nn.Linear(self.hidden_dim, self.ffn_dim, bias=False)
        self.down = nn.Linear(self.ffn_dim, self.hidden_dim, bias=False)
        
    def forward(self, secondary_gate_input: torch.Tensor, hidden_states: torch.Tensor):
        if self.moe_enable_secondary_experts:
            secondary_gate_logits = F.sigmoid(self.secondary_gate(secondary_gate_input)) > 0.5
            secondary_gate_mask = secondary_gate_logits.unsqueeze(-1)

        current_hidden_states = self.up(hidden_states) * F.relu(self.gate(hidden_states))
        activated_output =  current_hidden_states
        batch_size, intermediate_size = activated_output.shape

        if self.moe_enable_secondary_experts:
            num_groups = intermediate_size // self.secondary_expert_size
            activated_output = activated_output.view(batch_size, num_groups, self.secondary_expert_size)
            output = activated_output * secondary_gate_mask
        else:
            output = activated_output

        current_hidden_states = output.view(batch_size, -1)
        current_hidden_states = self.down(current_hidden_states)
        return current_hidden_states


class SmallthinkerMoeBlock(nn.Module):
    def __init__(self, config: SmallthinkerConfig):
        super().__init__()
        self.hidden_dim = config.hidden_size
        self.num_primary_experts = config.moe_num_primary_experts
        self.enable_early_router = config.moe_enable_early_router
        self.moe_primary_router_apply_softmax = config.moe_primary_router_apply_softmax
        self.num_active_primary_experts = config.moe_num_active_primary_experts
        self.primary_router = nn.Linear(self.hidden_dim, self.num_primary_experts, bias=False)
        self.experts = nn.ModuleList([SmallthinkerHierarchicalMLP(config) for _ in range(self.num_primary_experts)])

    def forward(self, router_input: torch.Tensor, hidden_states: torch.Tensor) -> torch.Tensor:
        batch_size, sequence_length, hidden_dim = hidden_states.shape
        # Flatten the tokens into (bs * sl, hidden_dim)
        hidden_states = hidden_states.view(-1, hidden_dim)
        router_input = router_input.view(-1, hidden_dim)
        # Primary router logits: (bs * sl, n_experts)
        if self.enable_early_router:
            router_logits = self.primary_router(router_input)
        else:
            router_logits = self.primary_router(hidden_states)

        router_logits, selected_experts = torch.topk(router_logits, self.num_active_primary_experts, dim=-1)

        if self.moe_primary_router_apply_softmax:
            routing_weights = F.softmax(router_logits, dim=1, dtype=torch.float)
        else:
            routing_weights = F.sigmoid(router_logits)
            routing_weights /= routing_weights.sum(dim=-1, keepdim=True)

        routing_weights = routing_weights.to(hidden_states.dtype)

        # Prepare the final tensor
        final_hidden_states = torch.zeros(
            (batch_size * sequence_length, hidden_dim), dtype=hidden_states.dtype, device=hidden_states.device
        )

        # One hot encode the selected experts to create an expert mask
        # this will be used to easily index which expert is going to be sollicitated
        expert_mask = torch.nn.functional.one_hot(selected_experts, num_classes=self.num_primary_experts).permute(2, 1, 0)
        expert_hitted = (expert_mask.sum(dim=(-1, -2)) > 0).nonzero(as_tuple=True)[0].tolist()

        for expert_idx in expert_hitted:
            expert_layer = self.experts[expert_idx]
            idx, top_x = torch.where(expert_mask[expert_idx])
            # Index the correct hidden states and compute the expert hidden state for
            # the current expert. We need to make sure to multiply the output hidden
            # states by `routing_weights` on the corresponding tokens (top-1 and top-2)
            # current_state = hidden_states[None, top_x].reshape(-1, hidden_dim)
            # current_router_input = router_input[None, top_x].reshape(-1, hidden_dim)
            current_state = hidden_states[top_x].reshape(-1, hidden_dim)
            current_router_input = router_input[top_x].reshape(-1, hidden_dim)
            current_hidden_states = expert_layer(current_router_input, current_state) * routing_weights[top_x, idx, None]

            # However `index_add_` only support torch tensors for indexing so we'll use the `top_x` tensor here.
            final_hidden_states.index_add_(0, top_x, current_hidden_states.to(hidden_states.dtype))
        final_hidden_states = final_hidden_states.reshape(batch_size, sequence_length, hidden_dim)
        return final_hidden_states, router_logits
    

class SmallthinkerDenseMlpBlock(nn.Module):
    def __init__(self, config: SmallthinkerConfig):
        super().__init__()
        hidden_dim = config.hidden_size
        ffn_dim = config.dense_ffn_hidden_size
        self.up = nn.Linear(hidden_dim, ffn_dim, bias=False)
        self.gate = nn.Linear(hidden_dim, ffn_dim, bias=False)
        self.down = nn.Linear(ffn_dim, hidden_dim, bias=False)

    # Offer unified interface for SmallthinkerMoeBlock and SmallthinkerDenseMlpBlock, though router_input is not used here
    def forward(self, router_input: torch.Tensor, hidden_states: torch.Tensor) -> torch.Tensor:
        current_hidden_states = self.up(hidden_states) * F.relu(self.gate(hidden_states))
        current_hidden_states = self.down(current_hidden_states)
        return current_hidden_states, None


class SmallthinkerRMSNorm(nn.Module):
    def __init__(self, hidden_size, eps=1e-6):
        """
        SmallthinkerRMSNorm is equivalent to T5LayerNorm
        """
        super().__init__()
        self.hidden_size = hidden_size
        self.weight = nn.Parameter(torch.ones(hidden_size))
        self.variance_epsilon = eps

    def forward(self, hidden_states):
        input_dtype = hidden_states.dtype
        hidden_states = hidden_states.to(torch.float32)
        variance = hidden_states.pow(2).mean(-1, keepdim=True)
        hidden_states = hidden_states * torch.rsqrt(variance + self.variance_epsilon)
        return self.weight * hidden_states.to(input_dtype)

    def extra_repr(self):
        return f"{tuple(self.weight.shape)}, eps={self.variance_epsilon}"


def rotate_half(x):
    """Rotates half the hidden dims of the input."""
    x1 = x[..., : x.shape[-1] // 2]
    x2 = x[..., x.shape[-1] // 2 :]
    return torch.cat((-x2, x1), dim=-1)


def apply_rotary_pos_emb(q, k, cos, sin, position_ids=None, unsqueeze_dim=1):
    """Applies Rotary Position Embedding to the query and key tensors.

    Args:
        q (`torch.Tensor`): The query tensor.
        k (`torch.Tensor`): The key tensor.
        cos (`torch.Tensor`): The cosine part of the rotary embedding.
        sin (`torch.Tensor`): The sine part of the rotary embedding.
        position_ids (`torch.Tensor`, *optional*):
            Deprecated and unused.
        unsqueeze_dim (`int`, *optional*, defaults to 1):
            The 'unsqueeze_dim' argument specifies the dimension along which to unsqueeze cos[position_ids] and
            sin[position_ids] so that they can be properly broadcasted to the dimensions of q and k. For example, note
            that cos[position_ids] and sin[position_ids] have the shape [batch_size, seq_len, head_dim]. Then, if q and
            k have the shape [batch_size, heads, seq_len, head_dim], then setting unsqueeze_dim=1 makes
            cos[position_ids] and sin[position_ids] broadcastable to the shapes of q and k. Similarly, if q and k have
            the shape [batch_size, seq_len, heads, head_dim], then set unsqueeze_dim=2.
    Returns:
        `tuple(torch.Tensor)` comprising of the query and key tensors rotated using the Rotary Position Embedding.
    """
    cos = cos.unsqueeze(unsqueeze_dim)
    sin = sin.unsqueeze(unsqueeze_dim)
    q_embed = (q * cos) + (rotate_half(q) * sin)
    k_embed = (k * cos) + (rotate_half(k) * sin)
    return q_embed, k_embed


def repeat_kv(hidden_states: torch.Tensor, n_rep: int) -> torch.Tensor:
    """
    This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch,
    num_key_value_heads, seqlen, head_dim) to (batch, num_attention_heads, seqlen, head_dim)
    """
    batch, num_key_value_heads, slen, head_dim = hidden_states.shape
    if n_rep == 1:
        return hidden_states
    hidden_states = hidden_states[:, :, None, :, :].expand(batch, num_key_value_heads, n_rep, slen, head_dim)
    return hidden_states.reshape(batch, num_key_value_heads * n_rep, slen, head_dim)


def eager_attention_forward(
    module: nn.Module,
    query: torch.Tensor,
    key: torch.Tensor,
    value: torch.Tensor,
    attention_mask: Optional[torch.Tensor],
    scaling: float,
    dropout: float = 0.0,
    **kwargs,
):
    key_states = repeat_kv(key, module.num_key_value_groups)
    value_states = repeat_kv(value, module.num_key_value_groups)

    attn_weights = torch.matmul(query, key_states.transpose(2, 3)) * scaling
    if attention_mask is not None:
        causal_mask = attention_mask[:, :, :, : key_states.shape[-2]]
        attn_weights = attn_weights + causal_mask

    attn_weights = nn.functional.softmax(attn_weights, dim=-1, dtype=torch.float32).to(query.dtype)
    attn_weights = nn.functional.dropout(attn_weights, p=dropout, training=module.training)
    attn_output = torch.matmul(attn_weights, value_states)
    attn_output = attn_output.transpose(1, 2).contiguous()

    return attn_output, attn_weights


class SmallthinkerAttention(nn.Module):
    def __init__(self, config: SmallthinkerConfig, layer_idx: int):
        super().__init__()
        self.config = config
        self.layer_idx = layer_idx # For KVCache management
        self.head_dim = config.head_dim
        self.num_attention_heads = config.num_attention_heads
        self.num_key_value_heads = config.num_key_value_heads
        self.num_key_value_groups = config.num_attention_heads // config.num_key_value_heads
        self.scaling = self.head_dim**-0.5
        self.is_causal = True
        self.q_proj = nn.Linear(config.hidden_size, config.num_attention_heads * self.head_dim, bias=False)
        self.k_proj = nn.Linear(config.hidden_size, config.num_key_value_heads * self.head_dim, bias=False)
        self.v_proj = nn.Linear(config.hidden_size, config.num_key_value_heads * self.head_dim, bias=False)
        self.o_proj = nn.Linear(config.num_attention_heads * self.head_dim, config.hidden_size, bias=False)
        self.sliding_window = config.sliding_window_size if config.sliding_window_layout[layer_idx] else None
        self.use_qk_norm = config.use_qk_norm

    def forward(
        self,
        hidden_states: torch.Tensor,
        position_embeddings: Tuple[torch.Tensor, torch.Tensor],
        attention_mask: Optional[torch.Tensor],
        past_key_value: Optional[Cache] = None,
        cache_position: Optional[torch.LongTensor] = None,
        **kwargs: Unpack[FlashAttentionKwargs],
    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
        if self.use_qk_norm:
            raise NotImplementedError("use_qk_norm is not implemented yet")

        input_shape = hidden_states.shape[:-1]
        hidden_shape = (*input_shape, -1, self.head_dim)

        query_states = self.q_proj(hidden_states).view(hidden_shape).transpose(1, 2)
        key_states = self.k_proj(hidden_states).view(hidden_shape).transpose(1, 2)
        value_states = self.v_proj(hidden_states).view(hidden_shape).transpose(1, 2)
        
        if position_embeddings:
            cos, sin = position_embeddings
            query_states, key_states = apply_rotary_pos_emb(query_states, key_states, cos, sin)
        else:
            cos, sin = None, None
        if past_key_value is not None:
            # sin and cos are specific to RoPE models; cache_position needed for the static cache
            cache_kwargs = {"sin": sin, "cos": cos, "cache_position": cache_position}
            key_states, value_states = past_key_value.update(key_states, value_states, self.layer_idx, cache_kwargs)

        attention_interface: Callable = eager_attention_forward

        if self.config._attn_implementation == "sdpa":
            raise NotImplementedError("SDPA impl is buggy for now. NEVER TRY TO USE IT.")

        if self.config._attn_implementation != "eager":
            if self.config._attn_implementation == "sdpa" and kwargs.get("output_attentions", False):
                logger.warning_once(
                    "`torch.nn.functional.scaled_dot_product_attention` does not support `output_attentions=True`. Falling back to "
                    'eager attention. This warning can be removed using the argument `attn_implementation="eager"` when loading the model.'
                )
            else:
                attention_interface = ALL_ATTENTION_FUNCTIONS[self.config._attn_implementation]

        attn_output, attn_weights = attention_interface(
            self,
            query_states,
            key_states,
            value_states,
            attention_mask,
            dropout=0.0,
            scaling=self.scaling,
            sliding_window=self.sliding_window,  # main diff with Llama
            **kwargs,
        )

        attn_output = attn_output.reshape(*input_shape, -1).contiguous()
        attn_output = self.o_proj(attn_output)
        return attn_output, attn_weights


class SmallthinkerDecoderLayer(nn.Module):
    def __init__(self, config: SmallthinkerConfig, layer_idx: int):
        super().__init__()
        self.hidden_size = config.hidden_size
        self.self_attn = SmallthinkerAttention(config, layer_idx)

        self.block_sparse_moe = SmallthinkerMoeBlock(config) if config.moe_layer_layout[layer_idx] else SmallthinkerDenseMlpBlock(config)
        self.input_layernorm = SmallthinkerRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
        self.post_attention_layernorm = SmallthinkerRMSNorm(config.hidden_size, eps=config.rms_norm_eps)

    def forward(
        self,
        hidden_states: torch.Tensor,
        attention_mask: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        past_key_value: Optional[Tuple[torch.Tensor]] = None,
        output_attentions: Optional[bool] = False,
        output_router_logits: Optional[bool] = False,
        use_cache: Optional[bool] = False,
        cache_position: Optional[torch.LongTensor] = None,
        position_embeddings: Optional[Tuple[torch.Tensor, torch.Tensor]] = None,  # necessary, but kept here for BC
        **kwargs: Unpack[FlashAttentionKwargs],
    ) -> Tuple[torch.FloatTensor, Optional[Tuple[torch.FloatTensor, torch.FloatTensor]]]:
        """
        Args:
            hidden_states (`torch.FloatTensor`): input to the layer of shape `(batch, seq_len, embed_dim)`
            attention_mask (`torch.FloatTensor`, *optional*): attention mask of size
                `(batch, sequence_length)` where padding elements are indicated by 0.
            past_key_value (`Tuple(torch.FloatTensor)`, *optional*): cached past key and value projection states
            output_attentions (`bool`, *optional*):
                Whether or not to return the attentions tensors of all attention layers. See `attentions` under
                returned tensors for more detail.
            output_router_logits (`bool`, *optional*):
                Whether or not to return the logits of all the routers. They are useful for computing the router loss, and
                should not be returned during inference.
            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.
            kwargs (`dict`, *optional*):
                Arbitrary kwargs to be ignored, used for FSDP and other methods that injects code
                into the model
        """
        # print(f"hidden states, shape {hidden_states.shape}: {hidden_states}") # debug print
        residual = hidden_states
        router_input = hidden_states
        hidden_states = self.input_layernorm(hidden_states)
        # Self Attention 
        hidden_states, self_attn_weights = self.self_attn(
            hidden_states=hidden_states,
            position_embeddings=position_embeddings,
            attention_mask=attention_mask,
            position_ids=position_ids,
            past_key_value=past_key_value,
            output_attentions=output_attentions,
            use_cache=use_cache,
            cache_position=cache_position,
            **kwargs,
        )
        hidden_states = residual + hidden_states

        # Fully Connected
        residual = hidden_states
        hidden_states = self.post_attention_layernorm(hidden_states)
        hidden_states, router_logits = self.block_sparse_moe(router_input, hidden_states)
        hidden_states = residual + hidden_states # SYNC after_moe_residual_value=hidden_states

        outputs = (hidden_states,)

        if output_attentions:
            outputs += (self_attn_weights,)

        if output_router_logits:
            outputs += (router_logits,)

        return outputs


class SmallthinkerRotaryEmbedding(nn.Module):
    def __init__(self, config: SmallthinkerConfig, device=None):
        super().__init__()
        # BC: "rope_type" was originally "type"
        if hasattr(config, "rope_scaling") and config.rope_scaling is not None:
            self.rope_type = config.rope_scaling.get("rope_type", config.rope_scaling.get("type"))
        else:
            self.rope_type = "default"
        self.max_seq_len_cached = config.max_position_embeddings
        self.original_max_seq_len = config.max_position_embeddings

        self.config = config
        self.rope_init_fn = ROPE_INIT_FUNCTIONS[self.rope_type]

        inv_freq, self.attention_scaling = self.rope_init_fn(self.config, device)
        self.register_buffer("inv_freq", inv_freq, persistent=False)
        self.original_inv_freq = self.inv_freq

    @torch.no_grad()
    @dynamic_rope_update  # power user: used with advanced RoPE types (e.g. dynamic rope)
    def forward(self, x, position_ids):
        inv_freq_expanded = self.inv_freq[None, :, None].float().expand(position_ids.shape[0], -1, 1).to(x.device)
        position_ids_expanded = position_ids[:, None, :].float()

        device_type = x.device.type if isinstance(x.device.type, str) and x.device.type != "mps" else "cpu"
        with torch.autocast(device_type=device_type, enabled=False):  # Force float32
            freqs = (inv_freq_expanded.float() @ position_ids_expanded.float()).transpose(1, 2)
            emb = torch.cat((freqs, freqs), dim=-1)
            cos = emb.cos() * self.attention_scaling
            sin = emb.sin() * self.attention_scaling

        return cos.to(dtype=x.dtype), sin.to(dtype=x.dtype)


class SmallthinkerPreTrainedModel(PreTrainedModel):
    config_class = SmallthinkerConfig
    base_model_prefix = "model"
    supports_gradient_checkpointing = True
    _no_split_modules = ["SmallthinkerDecoderLayer"]
    _skip_keys_device_placement = ["past_key_values"]
    _supports_flash_attn_2 = True
    _supports_sdpa = True
    _supports_flex_attn = True
    _supports_cache_class = True
    _supports_quantized_cache = True
    _supports_static_cache = False  # MoE models don't work with torch.compile (`torch.where(condition)` not supported)
    _supports_attention_backend = True

    def _init_weights(self, module):
        std = self.config.initializer_range
        if isinstance(module, nn.Linear):
            module.weight.data.normal_(mean=0.0, std=std)
            if module.bias is not None:
                module.bias.data.zero_()
        elif isinstance(module, nn.Embedding):
            module.weight.data.normal_(mean=0.0, std=std)
            if module.padding_idx is not None:
                module.weight.data[module.padding_idx].zero_()
        elif isinstance(module, SmallthinkerRMSNorm):
            module.weight.data.fill_(1.0)


# @auto_docstring
class SmallthinkerModel(SmallthinkerPreTrainedModel):
    def __init__(self, config: SmallthinkerConfig):
        super().__init__(config)
        self.padding_idx = config.pad_token_id
        self.vocab_size = config.vocab_size

        self.embed_tokens = nn.Embedding(config.vocab_size, config.hidden_size, self.padding_idx)
        self.layers = nn.ModuleList(
            [SmallthinkerDecoderLayer(config, layer_idx) for layer_idx in range(config.num_hidden_layers)]
        )
        self.norm = SmallthinkerRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
        self.rotary_emb = SmallthinkerRotaryEmbedding(config=config)
        self.gradient_checkpointing = False
        self.rope_layout = config.rope_layout

        # Initialize weights and apply final processing
        self.post_init()

    def get_input_embeddings(self):
        return self.embed_tokens

    def set_input_embeddings(self, value):
        self.embed_tokens = value

    @can_return_tuple
    # @auto_docstring
    def forward(
        self,
        input_ids: Optional[torch.LongTensor] = None,
        attention_mask: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        past_key_values: Optional[List[torch.FloatTensor]] = None,
        inputs_embeds: Optional[torch.FloatTensor] = None,
        use_cache: Optional[bool] = None,
        output_attentions: Optional[bool] = None,
        output_hidden_states: Optional[bool] = None,
        output_router_logits: Optional[bool] = None,
        cache_position: Optional[torch.LongTensor] = None,
        **flash_attn_kwargs: Unpack[FlashAttentionKwargs],
    ) -> MoeModelOutputWithPast:
        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
        output_router_logits = (
            output_router_logits if output_router_logits is not None else self.config.output_router_logits
        )
        output_hidden_states = (
            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
        )
        use_cache = use_cache if use_cache is not None else self.config.use_cache

        if (input_ids is None) ^ (inputs_embeds is not None):
            raise ValueError("You must specify exactly one of input_ids or inputs_embeds")

        if self.gradient_checkpointing and self.training:
            if use_cache:
                logger.warning_once(
                    "`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`..."
                )
                use_cache = False

        if use_cache and past_key_values is None:
            past_key_values = DynamicCache()

        if inputs_embeds is None:
            inputs_embeds = self.embed_tokens(input_ids)

        if cache_position is None:
            past_seen_tokens = past_key_values.get_seq_length() if past_key_values is not None else 0
            cache_position = torch.arange(
                past_seen_tokens, past_seen_tokens + inputs_embeds.shape[1], device=inputs_embeds.device
            )
        if position_ids is None:
            position_ids = cache_position.unsqueeze(0)

        # print("atten mask:", attention_mask) # debug print

        causal_mask = self._update_causal_mask(
            attention_mask, inputs_embeds, cache_position, past_key_values, output_attentions
        )

        # print("causal mask:", causal_mask) # debug print
        hidden_states = inputs_embeds
        # create position embeddings to be shared across the decoder layers
        position_embeddings = self.rotary_emb(hidden_states, position_ids)

        # decoder layers
        all_hidden_states = () if output_hidden_states else None
        all_self_attns = () if output_attentions else None
        all_router_logits = () if output_router_logits else None

        for layer_idx, decoder_layer in enumerate(self.layers):
            if output_hidden_states:
                all_hidden_states += (hidden_states,)

            if self.gradient_checkpointing and self.training:
                layer_outputs = self._gradient_checkpointing_func(
                    partial(decoder_layer.__call__, **flash_attn_kwargs),
                    hidden_states,
                    causal_mask,
                    position_ids,
                    past_key_values,
                    output_attentions,
                    output_router_logits,
                    use_cache,
                    cache_position,
                    position_embeddings if self.rope_layout[layer_idx] else None,
                )
            else:
                layer_outputs = decoder_layer(
                    hidden_states,
                    attention_mask=causal_mask,
                    position_ids=position_ids,
                    past_key_value=past_key_values,
                    output_attentions=output_attentions,
                    output_router_logits=output_router_logits,
                    use_cache=use_cache,
                    cache_position=cache_position,
                    position_embeddings=position_embeddings if self.rope_layout[layer_idx] else None,
                    **flash_attn_kwargs,
                )

            hidden_states = layer_outputs[0]

            if output_attentions:
                all_self_attns += (layer_outputs[1],)

            if output_router_logits:
                all_router_logits += (layer_outputs[-1],)

        hidden_states = self.norm(hidden_states)

        # add hidden states from the last decoder layer
        if output_hidden_states:
            all_hidden_states += (hidden_states,)

        return MoeModelOutputWithPast(
            last_hidden_state=hidden_states,
            past_key_values=past_key_values,
            hidden_states=all_hidden_states,
            attentions=all_self_attns,
            router_logits=all_router_logits,
        )

    def _update_causal_mask(
        self,
        attention_mask: Union[torch.Tensor, "BlockMask"],
        input_tensor: torch.Tensor,
        cache_position: torch.Tensor,
        past_key_values: Cache,
        output_attentions: bool = False,
    ):
        if self.config._attn_implementation == "flash_attention_2":
            if attention_mask is not None and past_key_values is not None:
                is_padding_right = attention_mask[:, -1].sum().item() != input_tensor.size()[0]
                if is_padding_right:
                    raise ValueError(
                        "You are attempting to perform batched generation with padding_side='right'"
                        " this may lead to unexpected behaviour for Flash Attention version of Smallthinker. Make sure to "
                        " call `tokenizer.padding_side  = 'left'` before tokenizing the input. "
                    )
            if attention_mask is not None and 0.0 in attention_mask:
                return attention_mask
            return None
        if self.config._attn_implementation == "flex_attention":
            if isinstance(attention_mask, torch.Tensor):
                attention_mask = make_flex_block_causal_mask(attention_mask)
            return attention_mask

        # For SDPA, when possible, we will rely on its `is_causal` argument instead of its `attn_mask` argument, in
        # order to dispatch on Flash Attention 2. This feature is not compatible with static cache, as SDPA will fail
        # to infer the attention mask.
        past_seen_tokens = past_key_values.get_seq_length() if past_key_values is not None else 0
        using_static_cache = isinstance(past_key_values, StaticCache)
        using_sliding_window_cache = isinstance(past_key_values, SlidingWindowCache)

        # When output attentions is True, sdpa implementation's forward method calls the eager implementation's forward
        if (
            self.config._attn_implementation == "sdpa"
            and not (using_static_cache or using_sliding_window_cache)
            and not output_attentions
        ):
            if AttentionMaskConverter._ignore_causal_mask_sdpa(
                attention_mask,
                inputs_embeds=input_tensor,
                past_key_values_length=past_seen_tokens,
                sliding_window=self.config.sliding_window,
                is_training=self.training,
            ):
                return None

        dtype = input_tensor.dtype
        min_dtype = torch.finfo(dtype).min
        sequence_length = input_tensor.shape[1]
        # SlidingWindowCache or StaticCache
        if using_sliding_window_cache or using_static_cache:
            target_length = past_key_values.get_max_cache_shape()
        # DynamicCache or no cache
        else:
            target_length = (
                attention_mask.shape[-1]
                if isinstance(attention_mask, torch.Tensor)
                else past_seen_tokens + sequence_length + 1
            )

        # In case the provided `attention` mask is 2D, we generate a causal mask here (4D).
        causal_mask = self._prepare_4d_causal_attention_mask_with_cache_position(
            attention_mask,
            sequence_length=sequence_length,
            target_length=target_length,
            dtype=dtype,
            cache_position=cache_position,
            batch_size=input_tensor.shape[0],
            config=self.config,
            past_key_values=past_key_values,
        )

        if (
            self.config._attn_implementation == "sdpa"
            and attention_mask is not None
            and attention_mask.device.type in ["cuda", "xpu", "npu"]
            and not output_attentions
        ):
            # Attend to all tokens in fully masked rows in the causal_mask, for example the relevant first rows when
            # using left padding. This is required by F.scaled_dot_product_attention memory-efficient attention path.
            # Details: https://github.com/pytorch/pytorch/issues/110213
            causal_mask = AttentionMaskConverter._unmask_unattended(causal_mask, min_dtype)

        return causal_mask

    @staticmethod
    def _prepare_4d_causal_attention_mask_with_cache_position(
        attention_mask: torch.Tensor,
        sequence_length: int,
        target_length: int,
        dtype: torch.dtype,
        cache_position: torch.Tensor,
        batch_size: int,
        config: SmallthinkerConfig,
        past_key_values: Cache,
    ):
        """
        Creates a causal 4D mask of shape `(batch_size, 1, query_length, key_value_length)` from a 2D mask of shape
        `(batch_size, key_value_length)`, or if the input `attention_mask` is already 4D, do nothing.

        Args:
            attention_mask (`torch.Tensor`):
                A 2D attention mask of shape `(batch_size, key_value_length)` or a 4D attention mask of shape `(batch_size, 1, query_length, key_value_length)`.
            sequence_length (`int`):
                The sequence length being processed.
            target_length (`int`):
                The target length: when generating with static cache, the mask should be as long as the static cache, to account for the 0 padding, the part of the cache that is not filled yet.
            dtype (`torch.dtype`):
                The dtype to use for the 4D attention mask.
            cache_position (`torch.Tensor`):
                Indices depicting the position of the input sequence tokens in the sequence.
            batch_size (`torch.Tensor`):
                Batch size.
            config (`SmallthinkerConfig`):
                The model's configuration class
            past_key_values (`Cache`):
                The cache class that is being used currently to generate
        """
        if attention_mask is not None and attention_mask.dim() == 4:
            # In this case we assume that the mask comes already in inverted form and requires no inversion or slicing.
            causal_mask = attention_mask
        else:
            min_dtype = torch.finfo(dtype).min
            causal_mask = torch.full(
                (sequence_length, target_length), fill_value=min_dtype, dtype=dtype, device=cache_position.device
            )
            diagonal_attend_mask = torch.arange(target_length, device=cache_position.device) > cache_position.reshape(
                -1, 1
            )
            if config.get_text_config().sliding_window is not None:
                # if we have sliding window, we should not attend to tokens beyond sliding window length, so we mask them out also
                # the check is needed to verify is current checkpoint was trained with sliding window or not
                if not isinstance(past_key_values, SlidingWindowCache) or sequence_length > target_length:
                    sliding_attend_mask = torch.arange(target_length, device=cache_position.device) <= (
                        cache_position.reshape(-1, 1) - config.get_text_config().sliding_window
                    )
                    diagonal_attend_mask.bitwise_or_(sliding_attend_mask)
            causal_mask *= diagonal_attend_mask
            causal_mask = causal_mask[None, None, :, :].expand(batch_size, 1, -1, -1)
            if attention_mask is not None:
                causal_mask = causal_mask.clone()  # copy to contiguous memory for in-place edit
                if attention_mask.shape[-1] > target_length:
                    attention_mask = attention_mask[:, :target_length]
                mask_length = attention_mask.shape[-1]
                padding_mask = causal_mask[:, :, :, :mask_length] + attention_mask[:, None, None, :].to(
                    causal_mask.device
                )
                padding_mask = padding_mask == 0
                causal_mask[:, :, :, :mask_length] = causal_mask[:, :, :, :mask_length].masked_fill(
                    padding_mask, min_dtype
                )
        return causal_mask


class KwargsForCausalLM(FlashAttentionKwargs, LossKwargs): ...


def load_balancing_loss_func(
    gate_logits: Union[torch.Tensor, Tuple[torch.Tensor], None],
    num_experts: Optional[int] = None,
    top_k=2,
    attention_mask: Optional[torch.Tensor] = None,
) -> Union[torch.Tensor, int]:
    r"""
    Computes auxiliary load balancing loss as in Switch Transformer - implemented in Pytorch.

    See Switch Transformer (https://arxiv.org/abs/2101.03961) for more details. This function implements the loss
    function presented in equations (4) - (6) of the paper. It aims at penalizing cases where the routing between
    experts is too unbalanced.

    Args:
        gate_logits:
            Logits from the `gate`, should be a tuple of model.config.num_hidden_layers tensors of
            shape [batch_size X sequence_length, num_experts].
        num_experts:
            Number of experts
        top_k:
            The number of experts to route per-token, can be also interpreted as the `top-k` routing
            parameter.
        attention_mask (`torch.Tensor`, *optional*):
            The attention_mask used in forward function
            shape [batch_size X sequence_length] if not None.

    Returns:
        The auxiliary loss.
    """
    if gate_logits is None or not isinstance(gate_logits, tuple):
        return 0

    if isinstance(gate_logits, tuple):
        compute_device = gate_logits[0].device
        concatenated_gate_logits = torch.cat([layer_gate.to(compute_device) for layer_gate in gate_logits], dim=0)

    routing_weights = torch.nn.functional.softmax(concatenated_gate_logits, dim=-1)

    _, selected_experts = torch.topk(routing_weights, top_k, dim=-1)

    expert_mask = torch.nn.functional.one_hot(selected_experts, num_experts)

    if attention_mask is None:
        # Compute the percentage of tokens routed to each experts
        tokens_per_expert = torch.mean(expert_mask.float(), dim=0)

        # Compute the average probability of routing to these experts
        router_prob_per_expert = torch.mean(routing_weights, dim=0)
    else:
        batch_size, sequence_length = attention_mask.shape
        num_hidden_layers = concatenated_gate_logits.shape[0] // (batch_size * sequence_length)

        # Compute the mask that masks all padding tokens as 0 with the same shape of expert_mask
        expert_attention_mask = (
            attention_mask[None, :, :, None, None]
            .expand((num_hidden_layers, batch_size, sequence_length, top_k, num_experts))
            .reshape(-1, top_k, num_experts)
            .to(compute_device)
        )

        # Compute the percentage of tokens routed to each experts
        tokens_per_expert = torch.sum(expert_mask.float() * expert_attention_mask, dim=0) / torch.sum(
            expert_attention_mask, dim=0
        )

        # Compute the mask that masks all padding tokens as 0 with the same shape of tokens_per_expert
        router_per_expert_attention_mask = (
            attention_mask[None, :, :, None]
            .expand((num_hidden_layers, batch_size, sequence_length, num_experts))
            .reshape(-1, num_experts)
            .to(compute_device)
        )

        # Compute the average probability of routing to these experts
        router_prob_per_expert = torch.sum(routing_weights * router_per_expert_attention_mask, dim=0) / torch.sum(
            router_per_expert_attention_mask, dim=0
        )

    overall_loss = torch.sum(tokens_per_expert * router_prob_per_expert.unsqueeze(0))
    return overall_loss * num_experts


# @auto_docstring
class SmallthinkerForCausalLM(SmallthinkerPreTrainedModel, GenerationMixin):
    _tied_weights_keys = ["lm_head.weight"]
    def __init__(self, config):
        super().__init__(config)
        self.model = SmallthinkerModel(config)
        self.vocab_size = config.vocab_size
        # Handle tie / untie word embeddings
        self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)        
        # self.num_experts = config.num_local_experts
        # self.num_experts_per_tok = config.num_experts_per_tok
        # Initialize weights and apply final processing
        self.post_init()

    def get_input_embeddings(self):
        return self.model.embed_tokens

    def set_input_embeddings(self, value):
        self.model.embed_tokens = value

    def get_output_embeddings(self):
        return self.lm_head

    def set_output_embeddings(self, new_embeddings):
        self.lm_head = new_embeddings

    def set_decoder(self, decoder):
        self.model = decoder

    def get_decoder(self):
        return self.model

    @can_return_tuple
#     @auto_docstring
    def forward(
        self,
        input_ids: Optional[torch.LongTensor] = None,
        attention_mask: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        past_key_values: Optional[List[torch.FloatTensor]] = None,
        inputs_embeds: Optional[torch.FloatTensor] = None,
        labels: Optional[torch.LongTensor] = None,
        use_cache: Optional[bool] = None,
        output_attentions: Optional[bool] = None,
        output_hidden_states: Optional[bool] = None,
        output_router_logits: Optional[bool] = None,
        cache_position: Optional[torch.LongTensor] = None,
        logits_to_keep: Union[int, torch.Tensor] = 0,
        **kwargs: Unpack[KwargsForCausalLM],
    ) -> MoeCausalLMOutputWithPast:
        r"""
        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]`.

        Example:

        ```python
        >>> from transformers import AutoTokenizer, SmallthinkerForCausalLM

        >>> model = SmallthinkerForCausalLM.from_pretrained("mistralai/Smallthinker-8x7B-v0.1")
        >>> tokenizer = AutoTokenizer.from_pretrained("mistralai/Smallthinker-8x7B-v0.1")

        >>> 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."
        ```"""

        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
        output_router_logits = (
            output_router_logits if output_router_logits is not None else self.config.output_router_logits
        )

        output_hidden_states = (
            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
        )

        # decoder outputs consists of (dec_features, layer_state, dec_hidden, dec_attn)
        outputs: MoeModelOutputWithPast = self.model(
            input_ids=input_ids,
            attention_mask=attention_mask,
            position_ids=position_ids,
            past_key_values=past_key_values,
            inputs_embeds=inputs_embeds,
            use_cache=use_cache,
            output_attentions=output_attentions,
            output_hidden_states=output_hidden_states,
            output_router_logits=output_router_logits,
            cache_position=cache_position,
            **kwargs,
        )

        hidden_states = outputs.last_hidden_state
        # Only compute necessary logits, and do not upcast them to float if we are not computing the loss
        slice_indices = slice(-logits_to_keep, None) if isinstance(logits_to_keep, int) else logits_to_keep
        logits = self.lm_head(hidden_states[:, slice_indices, :])

        loss = None
        if labels is not None:
            loss = self.loss_function(logits, labels, self.vocab_size, **kwargs)

        aux_loss = None
        if output_router_logits:
            aux_loss = load_balancing_loss_func(
                outputs.router_logits,
                self.num_experts,
                self.num_experts_per_tok,
                attention_mask,
            )
            if labels is not None:
                loss += self.router_aux_loss_coef * aux_loss.to(loss.device)  # make sure to reside in the same device

        return MoeCausalLMOutputWithPast(
            loss=loss,
            aux_loss=aux_loss,
            logits=logits,
            past_key_values=outputs.past_key_values,
            hidden_states=outputs.hidden_states,
            attentions=outputs.attentions,
            router_logits=outputs.router_logits,
        )

# No such functions for now
# #@auto_docstring(
#     custom_intro="""
#     The Smallthinker Model transformer with a sequence classification head on top (linear layer).

#     [`SmallthinkerForSequenceClassification`] uses the last token in order to do the classification, as other causal models
#     (e.g. GPT-2) do.

#     Since it does classification on the last token, it requires to know the position of the last token. If a
#     `pad_token_id` is defined in the configuration, it finds the last token that is not a padding token in each row. If
#     no `pad_token_id` is defined, it simply takes the last value in each row of the batch. Since it cannot guess the
#     padding tokens when `inputs_embeds` are passed instead of `input_ids`, it does the same (take the last value in
#     each row of the batch).
#     """
# )
# class SmallthinkerForSequenceClassification(SmallthinkerPreTrainedModel):
#     def __init__(self, config):
#         super().__init__(config)
#         self.num_labels = config.num_labels
#         self.model = SmallthinkerModel(config)
#         self.score = nn.Linear(config.hidden_size, self.num_labels, bias=False)

#         # Initialize weights and apply final processing
#         self.post_init()

#     def get_input_embeddings(self):
#         return self.model.embed_tokens

#     def set_input_embeddings(self, value):
#         self.model.embed_tokens = value

#     @can_return_tuple
#     #@auto_docstring
#     def forward(
#         self,
#         input_ids: Optional[torch.LongTensor] = None,
#         attention_mask: Optional[torch.Tensor] = None,
#         position_ids: Optional[torch.LongTensor] = None,
#         past_key_values: Optional[Cache] = None,
#         inputs_embeds: Optional[torch.FloatTensor] = None,
#         labels: Optional[torch.LongTensor] = None,
#         use_cache: Optional[bool] = None,
#         output_attentions: Optional[bool] = None,
#         output_hidden_states: Optional[bool] = None,
#     ) -> SequenceClassifierOutputWithPast:
#         r"""
#         labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
#             Labels for computing the sequence classification/regression loss. Indices should be in `[0, ...,
#             config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If
#             `config.num_labels > 1` a classification loss is computed (Cross-Entropy).
#         """

#         transformer_outputs: BaseModelOutputWithPast = self.model(
#             input_ids,
#             attention_mask=attention_mask,
#             position_ids=position_ids,
#             past_key_values=past_key_values,
#             inputs_embeds=inputs_embeds,
#             use_cache=use_cache,
#             output_attentions=output_attentions,
#             output_hidden_states=output_hidden_states,
#         )
#         hidden_states = transformer_outputs.last_hidden_state
#         logits = self.score(hidden_states)

#         if input_ids is not None:
#             batch_size = input_ids.shape[0]
#         else:
#             batch_size = inputs_embeds.shape[0]

#         if self.config.pad_token_id is None and batch_size != 1:
#             raise ValueError("Cannot handle batch sizes > 1 if no padding token is defined.")
#         if self.config.pad_token_id is None:
#             last_non_pad_token = -1
#         elif input_ids is not None:
#             # To handle both left- and right- padding, we take the rightmost token that is not equal to pad_token_id
#             non_pad_mask = (input_ids != self.config.pad_token_id).to(logits.device, torch.int32)
#             token_indices = torch.arange(input_ids.shape[-1], device=logits.device, dtype=torch.int32)
#             last_non_pad_token = (token_indices * non_pad_mask).argmax(-1)
#         else:
#             last_non_pad_token = -1
#             logger.warning_once(
#                 f"{self.__class__.__name__} will not detect padding tokens in `inputs_embeds`. Results may be "
#                 "unexpected if using padding tokens in conjunction with `inputs_embeds.`"
#             )

#         pooled_logits = logits[torch.arange(batch_size, device=logits.device), last_non_pad_token]

#         loss = None
#         if labels is not None:
#             loss = self.loss_function(logits=logits, labels=labels, pooled_logits=pooled_logits, config=self.config)

#         return SequenceClassifierOutputWithPast(
#             loss=loss,
#             logits=pooled_logits,
#             past_key_values=transformer_outputs.past_key_values,
#             hidden_states=transformer_outputs.hidden_states,
#             attentions=transformer_outputs.attentions,
#         )


# #@auto_docstring
# class SmallthinkerForTokenClassification(SmallthinkerPreTrainedModel):
#     def __init__(self, config):
#         super().__init__(config)
#         self.num_labels = config.num_labels
#         self.model = SmallthinkerModel(config)
#         if getattr(config, "classifier_dropout", None) is not None:
#             classifier_dropout = config.classifier_dropout
#         elif getattr(config, "hidden_dropout", None) is not None:
#             classifier_dropout = config.hidden_dropout
#         else:
#             classifier_dropout = 0.1
#         self.dropout = nn.Dropout(classifier_dropout)
#         self.score = nn.Linear(config.hidden_size, config.num_labels)

#         # Initialize weights and apply final processing
#         self.post_init()

#     def get_input_embeddings(self):
#         return self.model.embed_tokens

#     def set_input_embeddings(self, value):
#         self.model.embed_tokens = value

#     @can_return_tuple
#     #@auto_docstring
#     def forward(
#         self,
#         input_ids: Optional[torch.LongTensor] = None,
#         attention_mask: Optional[torch.Tensor] = None,
#         position_ids: Optional[torch.LongTensor] = None,
#         past_key_values: Optional[Cache] = None,
#         inputs_embeds: Optional[torch.FloatTensor] = None,
#         labels: Optional[torch.LongTensor] = None,
#         use_cache: Optional[bool] = None,
#         output_attentions: Optional[bool] = None,
#         output_hidden_states: Optional[bool] = None,
#     ) -> TokenClassifierOutput:
#         r"""
#         labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
#             Labels for computing the sequence classification/regression loss. Indices should be in `[0, ...,
#             config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If
#             `config.num_labels > 1` a classification loss is computed (Cross-Entropy).
#         """

#         outputs: BaseModelOutputWithPast = self.model(
#             input_ids,
#             attention_mask=attention_mask,
#             position_ids=position_ids,
#             past_key_values=past_key_values,
#             inputs_embeds=inputs_embeds,
#             use_cache=use_cache,
#             output_attentions=output_attentions,
#             output_hidden_states=output_hidden_states,
#         )
#         sequence_output = outputs.last_hidden_state
#         sequence_output = self.dropout(sequence_output)
#         logits = self.score(sequence_output)

#         loss = None
#         if labels is not None:
#             loss = self.loss_function(logits, labels, self.config)

#         return TokenClassifierOutput(
#             loss=loss,
#             logits=logits,
#             hidden_states=outputs.hidden_states,
#             attentions=outputs.attentions,
#         )


# #@auto_docstring
# class SmallthinkerForQuestionAnswering(SmallthinkerPreTrainedModel):
#     base_model_prefix = "model"

#     def __init__(self, config):
#         super().__init__(config)
#         self.qa_outputs = nn.Linear(config.hidden_size, 2)
#         self.model = SmallthinkerModel(config)  # diff with Llama: transformer->model

#         # Initialize weights and apply final processing
#         self.post_init()

#     def get_input_embeddings(self):
#         return self.model.embed_tokens

#     def set_input_embeddings(self, value):
#         self.model.embed_tokens = value

#     @can_return_tuple
#     #@auto_docstring
#     def forward(
#         self,
#         input_ids: Optional[torch.LongTensor] = None,
#         attention_mask: Optional[torch.Tensor] = None,
#         position_ids: Optional[torch.LongTensor] = None,
#         past_key_values: Optional[Union[Cache, List[torch.FloatTensor]]] = None,
#         inputs_embeds: Optional[torch.FloatTensor] = None,
#         start_positions: Optional[torch.LongTensor] = None,
#         end_positions: Optional[torch.LongTensor] = None,
#         output_attentions: Optional[bool] = None,
#         output_hidden_states: Optional[bool] = None,
#         **kwargs,
#     ) -> QuestionAnsweringModelOutput:
#         r"""
#         start_positions (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
#             Labels for position (index) of the start of the labelled span for computing the token classification loss.
#             Positions are clamped to the length of the sequence (`sequence_length`). Position outside of the sequence
#             are not taken into account for computing the loss.
#         end_positions (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
#             Labels for position (index) of the end of the labelled span for computing the token classification loss.
#             Positions are clamped to the length of the sequence (`sequence_length`). Position outside of the sequence
#             are not taken into account for computing the loss.
#         """

#         outputs: BaseModelOutputWithPast = self.model(
#             input_ids,
#             attention_mask=attention_mask,
#             position_ids=position_ids,
#             past_key_values=past_key_values,
#             inputs_embeds=inputs_embeds,
#             output_attentions=output_attentions,
#             output_hidden_states=output_hidden_states,
#         )

#         sequence_output = outputs.last_hidden_state

#         logits = self.qa_outputs(sequence_output)
#         start_logits, end_logits = logits.split(1, dim=-1)
#         start_logits = start_logits.squeeze(-1).contiguous()
#         end_logits = end_logits.squeeze(-1).contiguous()

#         loss = None
#         if start_positions is not None and end_positions is not None:
#             loss = self.loss_function(start_logits, end_logits, start_positions, end_positions, **kwargs)

#         return QuestionAnsweringModelOutput(
#             loss=loss,
#             start_logits=start_logits,
#             end_logits=end_logits,
#             hidden_states=outputs.hidden_states,
#             attentions=outputs.attentions,
#         )


__all__ = [
    "SmallthinkerForCausalLM",
    "SmallthinkerForQuestionAnswering",
    "SmallthinkerModel",
    "SmallthinkerPreTrainedModel",
    "SmallthinkerForSequenceClassification",
    "SmallthinkerForTokenClassification",
]

if __name__ == "__main__":
    from transformers import AutoTokenizer, AutoModelForCausalLM

    test_config = SmallthinkerConfig()
    tokenizer = AutoTokenizer.from_pretrained("./qwen-tokenizer")
    text = "Once upon a day"
    tokens = tokenizer.encode_plus( text,add_special_tokens=True,return_tensors='pt')
    # print(tokens)
    test_model = AutoModelForCausalLM.from_pretrained(".").cuda()

    output = test_model.generate(tokens)
    otokens = tokenizer.decode(output[0])
    # print(otokens)