Add files using upload-large-folder tool

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

chat_template.jinja

@@ -0,0 +1,103 @@
+[gMASK]<sop>
+{%- if tools -%}
+<|system|>
+# Tools
+
+You may call one or more functions to assist with the user query.
+
+You are provided with function signatures within <tools></tools> XML tags:
+<tools>
+{% for tool in tools %}
+{{ tool | tojson(ensure_ascii=False) }}
+{% endfor %}
+</tools>
+
+For each function call, output the function name and arguments within the following XML format:
+<tool_call>{function-name}
+<arg_key>{arg-key-1}</arg_key>
+<arg_value>{arg-value-1}</arg_value>
+<arg_key>{arg-key-2}</arg_key>
+<arg_value>{arg-value-2}</arg_value>
+...
+</tool_call>{%- endif -%}
+{%- macro visible_text(content) -%}
+    {%- if content is string -%}
+        {{- content }}
+    {%- elif content is iterable and content is not mapping -%}
+        {%- for item in content -%}
+            {%- if item is mapping and item.type == 'text' -%}
+                {{- item.text }}
+            {%- elif item is string -%}
+                {{- item }}
+            {%- endif -%}
+        {%- endfor -%}
+    {%- else -%}
+        {{- content }}
+    {%- endif -%}
+{%- endmacro -%}
+{%- set ns = namespace(last_user_index=-1) %}
+{%- for m in messages %}
+    {%- if m.role == 'user' %}
+        {% set ns.last_user_index = loop.index0 -%}
+    {%- endif %}
+{%- endfor %}
+{% for m in messages %}
+{%- if m.role == 'user' -%}<|user|>
+{{ visible_text(m.content) }}
+{{- '/nothink' if (enable_thinking is defined and not enable_thinking and not visible_text(m.content).endswith("/nothink")) else '' -}}
+{%- elif m.role == 'assistant' -%}
+<|assistant|>
+{%- set reasoning_content = '' %}
+{%- set content = visible_text(m.content) %}
+{%- if m.reasoning_content is string %}
+    {%- set reasoning_content = m.reasoning_content %}
+{%- else %}
+    {%- if '</think>' in content %}
+        {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+        {%- set content = content.split('</think>')[-1].lstrip('\n') %}
+    {%- endif %}
+{%- endif %}
+{%- if loop.index0 > ns.last_user_index and reasoning_content -%}
+{{ '\n<think>' + reasoning_content.strip() +  '</think>'}}
+{%- else -%}
+{{ '\n<think></think>' }}
+{%- endif -%}
+{%- if content.strip() -%}
+{{ '\n' + content.strip() }}
+{%- endif -%}
+{% if m.tool_calls %}
+{% for tc in m.tool_calls %}
+{%- if tc.function %}
+    {%- set tc = tc.function %}
+{%- endif %}
+{{ '\n<tool_call>' + tc.name }}
+{% set _args = tc.arguments %}
+{% for k, v in _args.items() %}
+<arg_key>{{ k }}</arg_key>
+<arg_value>{{ v | tojson(ensure_ascii=False) if v is not string else v }}</arg_value>
+{% endfor %}
+</tool_call>{% endfor %}
+{% endif %}
+{%- elif m.role == 'tool' -%}
+{%- if m.content is string -%}
+{%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+    {{- '<|observation|>' }}
+{%- endif %}
+{{- '\n<tool_response>\n' }}
+{{- m.content }}
+{{- '\n</tool_response>' }}
+{%- else -%}
+<|observation|>{% for tr in m.content %}
+
+<tool_response>
+{{ tr.output if tr.output is defined else tr }}
+</tool_response>{% endfor -%}
+{% endif -%}
+{%- elif m.role == 'system' -%}
+<|system|>
+{{ visible_text(m.content) }}
+{%- endif -%}
+{%- endfor -%}
+{%- if add_generation_prompt -%}
+    <|assistant|>{{- '\n<think></think>' if (enable_thinking is defined and not enable_thinking) else '' -}}
+{%- endif -%}

config.json

@@ -0,0 +1,46 @@
+{
+  "architectures": [
+    "Glm4MoeForCausalLM"
+  ],
+  "attention_bias": true,
+  "attention_dropout": 0.0,
+  "auto_map": {
+    "AutoModelForCausalLM": "modeling_glm4_moe.Glm4MoeForCausalLM"
+  },
+  "eos_token_id": [
+    151329,
+    151336,
+    151338
+  ],
+  "first_k_dense_replace": 3,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 5120,
+  "initializer_range": 0.02,
+  "intermediate_size": 12288,
+  "max_position_embeddings": 202752,
+  "model_type": "glm4_moe",
+  "moe_intermediate_size": 1536,
+  "n_group": 1,
+  "n_routed_experts": 80,
+  "n_shared_experts": 1,
+  "norm_topk_prob": true,
+  "num_attention_heads": 96,
+  "num_experts_per_tok": 8,
+  "num_hidden_layers": 92,
+  "num_key_value_heads": 8,
+  "num_nextn_predict_layers": 1,
+  "pad_token_id": 151329,
+  "partial_rotary_factor": 0.5,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": null,
+  "rope_theta": 1000000,
+  "routed_scaling_factor": 2.5,
+  "tie_word_embeddings": false,
+  "topk_group": 1,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.55.0",
+  "use_cache": true,
+  "use_qk_norm": true,
+  "vocab_size": 151552
+}

generation_config.json

@@ -0,0 +1,10 @@
+{
+  "_from_model_config": true,
+  "eos_token_id": [
+    151329,
+    151336,
+    151338
+  ],
+  "pad_token_id": 151329,
+  "transformers_version": "4.55.0"
+}

model-00001-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dc2614aad7b7f125cbe6665005daf458deeb2415c4125ee5d1b49c7601676d14
+size 4986172552

model-00004-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:493b79e8398d37f09af8ed27b9c86a5262412b106de57843aac55669531b7784
+size 4914326208

model-00005-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1183507706c9b62bc5c6cb21d50c9e35d458677be5134adbce802e054131355a
+size 4997400520

model-00007-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0aa881e1ad09d159742c3e1a3e83d2c532b7017a236e6479fd5a8a3feae3a11e
+size 4992129752

model-00008-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d89035f2d7cfa55395baadcb6d9af4fe75803566cac1924044de4cd68e393bfb
+size 4992129888

model-00010-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1aaa899b941123f413c99515d60737d9204bcf9027fc71a29e44d1f9be3a5401
+size 4992129872

model-00016-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:639ea6c4185edfb1fafd584333a17217c469f357239b6579ad127167e52a137c
+size 4992129888

model-00020-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b47f9c7e0a956ceae7b24c6f40f49a3893ceb6b67fea265e0019234546789aa6
+size 4992129888

model-00022-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5fd00ebf2d982f3953188a7fc07858f2ebc1caea9f3156bfd19e7d5ff30782f0
+size 4992129888

model-00023-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:102c92f4f24c856fff2b9bbb8061b4a8845b0b845e0491d82385059f38db340f
+size 4998241280

model-00024-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5b66f97680b69f674faa733cadd93681989cf821b5a1abe3671df82e16369133
+size 4992129888

model-00025-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c47cf41434b3bae0b12f2368f1108b02eb47600cdd8e8d8f41fff05e76be933c
+size 4992129888

model-00026-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:270868fb827dcd0bc71383fc310955125461ef87954a35661feb107a4f71ff17
+size 4992129888

model-00027-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:50938aa2a5ad56baa0dd9799632e9f4e169ab71dee59046cf68694a47379e7a3
+size 4966783928

model-00028-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2649719678c0bff61e4e46fc51c96af3140f173db2187d5a2f3b2d20838ec829
+size 4992129712

model-00032-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a541c7503038db458ee565899ed2c762502d877cb21dfc68cb83ae51817d9897
+size 4998241288

model-00035-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:79d9054b9d5008ce549c20f47fce85b20c5fc20fba97e27fd1e931f41b6bac17
+size 4992129888

model-00037-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a8f1a9dd3c6482fac33eee59474a07d34dc47c33a43136aa01e46ea56ee4252d
+size 4997400824

model-00038-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:70ebe15ead11ae38b3201bf5804bc2610afa277cc4a751447163e57d6c6640e7
+size 4992129888

model-00039-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6c130a9b072c765aaec3426883e4e5588ca991511637ff7317013386876259f6
+size 4992129888

model-00041-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f3e8959a0eaaa8c84e323ca03f33e85c4b42c63a5fe861d32f9e7f716a73b994
+size 4998241296

model-00042-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ef9f8ff060a8e72f12cc73ff390ae2e1689c039c4d7ad25a8933935bd8afdd32
+size 4992129872

model-00043-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:db490d5668c50934aeadf4bddca146658f2292cba09aafd69178a6ebfafffa35
+size 4992129888

model-00047-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:199b559733ae5a1a9130c0b63a54d5a52dd8896c5dfddf192dd102e6c56671df
+size 4992129888

model-00050-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:88bc64b28d732b669e9a99daf7d6a9ba4146b9dfb63da21609e97c491d26b0fc
+size 4998241304

model-00052-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c4b7f8275bf25321fb34015ef0e4f064068898a3408b4d64dd218dac7b59f63
+size 4992129888

model-00054-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e9155526068ed0b07761b84d613f9df11759c8778955d3c0cab8dddf5e513bb
+size 4992129888

model-00055-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:60e7adb20158d493162cd8d19de5c36b92c7a1fd1a2b5a61a77b87dd32b1fc8b
+size 4998241280

model-00056-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e0ec8f6944758593ecd406a1538962a81ab251580273955b7caa7a6d72ae00f
+size 4992129888

model-00059-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:40332b2c2e3073b3eadaca092cb06e85a5be36856cd3dc89ce8e13fa316f7b3d
+size 4966783928

model-00060-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bd800f47b5d18ec7a0eb7833a57087a1d5c0877ea52e62a81960c5f585ed7ab6
+size 4992129712

model-00061-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7216dd3f84d8817efcc0369c9a2e996105b4a56a04babc067cb188d53d5c303d
+size 4992129888

model-00064-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fbcaa525dd438741cf4edde64edad984d535118cfb226f2d1df584664c118b87
+size 4998241288

model-00066-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8578888eb758d57317b1763e61284e302f62697fae64dbba47c7f44a5cff7eee
+size 4992129888

model-00067-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4dbf39269b4bdec2d96c0f5cb128fef777288e18b76fdc24e9f8d15316d343e5
+size 4992129888

model-00068-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fd6da0d4f40e5652ced325a888746cf48caa17d65ee67bfa8c2ffe2a3b0e3987
+size 4961512816

model-00069-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dbbb42468a112a1b466ec12a2e371c41f87a5248d1e87ede340d48f11a3f4363
+size 4997400824

model-00071-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c2b09a61b9db2866fe55e000a9142b4c6f4c8e2042c92a22914f366ab5955334
+size 4992129888

model-00072-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:755b0f2c87b75ff87d0768f86ed0f60f952222a3abd335978567cbd136c45774
+size 4992129888

model-00074-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77b1a60271e8afe5762f608bab4edd3e0b1a40e3c8234065744e13e480189b11
+size 3697110504

model-00075-of-00075.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7ed732a1f52d39814adc6a0b127bad1c816b89d7055fa07b029cba98deba7bf
+size 1551892608

modeling_glm4_moe.py

@@ -0,0 +1,624 @@
+#                🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨
+#           This file was automatically generated from src/transformers/models/glm4_moe/modular_glm4_moe.py.
+#               Do NOT edit this file manually as any edits will be overwritten by the generation of
+#             the file from the modular. If any change should be done, please apply the change to the
+#                          modular_glm4_moe.py file directly. One of our CI enforces this.
+#                🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨
+# coding=utf-8
+# Copyright 2025 The ZhipuAI Inc. team and HuggingFace Inc. team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from typing import Callable, Optional, Union
+
+import torch
+import torch.nn.functional as F
+from torch import nn
+
+from transformers.activations import ACT2FN
+from transformers.cache_utils import Cache, DynamicCache
+from transformers.generation import GenerationMixin
+from transformers.integrations import use_kernel_forward_from_hub
+from transformers.masking_utils import create_causal_mask
+from transformers.modeling_flash_attention_utils import FlashAttentionKwargs
+from transformers.modeling_layers import GradientCheckpointingLayer
+from transformers.modeling_outputs import BaseModelOutputWithPast, CausalLMOutputWithPast
+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 TransformersKwargs, auto_docstring, can_return_tuple
+from transformers.utils.generic import check_model_inputs
+from transformers.models.glm4_moe.configuration_glm4_moe import Glm4MoeConfig
+
+
+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: Unpack[TransformersKwargs],
+):
+    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
+
+
+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)
+
+    # Keep half or full tensor for later concatenation
+    rotary_dim = cos.shape[-1]
+    q_rot, q_pass = q[..., :rotary_dim], q[..., rotary_dim:]
+    k_rot, k_pass = k[..., :rotary_dim], k[..., rotary_dim:]
+
+    # Apply rotary embeddings on the first half or full tensor
+    q_embed = (q_rot * cos) + (rotate_half(q_rot) * sin)
+    k_embed = (k_rot * cos) + (rotate_half(k_rot) * sin)
+
+    # Concatenate back to full shape
+    q_embed = torch.cat([q_embed, q_pass], dim=-1)
+    k_embed = torch.cat([k_embed, k_pass], dim=-1)
+    return q_embed, k_embed
+
+
+class Glm4MoeAttention(nn.Module):
+    """Multi-headed attention from 'Attention Is All You Need' paper"""
+
+    def __init__(self, config: Glm4MoeConfig, layer_idx: Optional[int] = None):
+        super().__init__()
+        self.config = config
+        self.layer_idx = layer_idx
+        self.head_dim = getattr(config, "head_dim", config.hidden_size // config.num_attention_heads)
+        self.num_key_value_groups = config.num_attention_heads // config.num_key_value_heads
+        self.scaling = self.head_dim**-0.5
+        self.attention_dropout = config.attention_dropout
+        self.is_causal = True
+
+        self.q_proj = nn.Linear(
+            config.hidden_size, config.num_attention_heads * self.head_dim, bias=config.attention_bias
+        )
+        self.k_proj = nn.Linear(
+            config.hidden_size, config.num_key_value_heads * self.head_dim, bias=config.attention_bias
+        )
+        self.v_proj = nn.Linear(
+            config.hidden_size, config.num_key_value_heads * self.head_dim, bias=config.attention_bias
+        )
+        self.o_proj = nn.Linear(config.num_attention_heads * self.head_dim, config.hidden_size, bias=False)
+        self.use_qk_norm = config.use_qk_norm
+        if self.use_qk_norm:
+            self.q_norm = Glm4MoeRMSNorm(self.head_dim, eps=config.rms_norm_eps)
+            self.k_norm = Glm4MoeRMSNorm(self.head_dim, eps=config.rms_norm_eps)
+
+    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]]:
+        input_shape = hidden_states.shape[:-1]
+        hidden_shape = (*input_shape, -1, self.head_dim)
+
+        query_states = self.q_proj(hidden_states).view(hidden_shape)
+        key_states = self.k_proj(hidden_states).view(hidden_shape)
+        value_states = self.v_proj(hidden_states).view(hidden_shape)
+
+        if self.use_qk_norm:  # main diff from Llama
+            query_states = self.q_norm(query_states)
+            key_states = self.k_norm(key_states)
+
+        query_states = query_states.transpose(1, 2)
+        key_states = key_states.transpose(1, 2)
+        value_states = value_states.transpose(1, 2)
+
+        cos, sin = position_embeddings
+        query_states, key_states = apply_rotary_pos_emb(query_states, key_states, cos, sin)
+
+        if past_key_value is not None:
+            # sin and cos are specific to RoPE models; position_ids 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 != "eager":
+            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 if not self.training else self.attention_dropout,
+            scaling=self.scaling,
+            **kwargs,
+        )
+
+        attn_output = attn_output.reshape(*input_shape, -1).contiguous()
+        attn_output = self.o_proj(attn_output)
+        return attn_output, attn_weights
+
+
+class Glm4MoeMLP(nn.Module):
+    def __init__(self, config, hidden_size=None, intermediate_size=None):
+        super().__init__()
+        self.config = config
+        self.hidden_size = config.hidden_size if hidden_size is None else hidden_size
+        self.intermediate_size = config.intermediate_size if intermediate_size is None else intermediate_size
+
+        self.gate_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
+        self.up_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
+        self.down_proj = nn.Linear(self.intermediate_size, self.hidden_size, bias=False)
+        self.act_fn = ACT2FN[config.hidden_act]
+
+    def forward(self, x):
+        down_proj = self.down_proj(self.act_fn(self.gate_proj(x)) * self.up_proj(x))
+        return down_proj
+
+
+class Glm4MoeTopkRouter(nn.Module):
+    def __init__(self, config: Glm4MoeConfig):
+        super().__init__()
+        self.config = config
+        self.top_k = config.num_experts_per_tok
+        self.n_routed_experts = config.n_routed_experts
+        self.routed_scaling_factor = config.routed_scaling_factor
+        self.n_group = config.n_group
+        self.topk_group = config.topk_group
+        self.norm_topk_prob = config.norm_topk_prob
+
+        self.weight = nn.Parameter(torch.empty((self.n_routed_experts, config.hidden_size)))
+        self.register_buffer("e_score_correction_bias", torch.zeros((self.n_routed_experts), dtype=torch.float32))
+
+    @torch.no_grad()
+    def get_topk_indices(self, scores):
+        scores_for_choice = scores.view(-1, self.n_routed_experts) + self.e_score_correction_bias.unsqueeze(0)
+        group_scores = (
+            scores_for_choice.view(-1, self.n_group, self.n_routed_experts // self.n_group)
+            .topk(2, dim=-1)[0]
+            .sum(dim=-1)
+        )
+        group_idx = torch.topk(group_scores, k=self.topk_group, dim=-1, sorted=False)[1]
+        group_mask = torch.zeros_like(group_scores)
+        group_mask.scatter_(1, group_idx, 1)
+        score_mask = (
+            group_mask.unsqueeze(-1)
+            .expand(-1, self.n_group, self.n_routed_experts // self.n_group)
+            .reshape(-1, self.n_routed_experts)
+        )
+        scores_for_choice = scores_for_choice.masked_fill(~score_mask.bool(), 0.0)
+        topk_indices = torch.topk(scores_for_choice, k=self.top_k, dim=-1, sorted=False)[1]
+        return topk_indices
+
+    def forward(self, hidden_states):
+        hidden_states = hidden_states.view(-1, self.config.hidden_size)
+        router_logits = F.linear(hidden_states.type(torch.float32), self.weight.type(torch.float32))
+        scores = router_logits.sigmoid()
+        topk_indices = self.get_topk_indices(scores)
+        topk_weights = scores.gather(1, topk_indices)
+        if self.norm_topk_prob:
+            denominator = topk_weights.sum(dim=-1, keepdim=True) + 1e-20
+            topk_weights /= denominator
+        topk_weights = topk_weights * self.routed_scaling_factor
+        return topk_indices, topk_weights, router_logits
+
+
+@use_kernel_forward_from_hub("RMSNorm")
+class Glm4MoeRMSNorm(nn.Module):
+    def __init__(self, hidden_size, eps=1e-6):
+        """
+        Glm4MoeRMSNorm is equivalent to T5LayerNorm
+        """
+        super().__init__()
+        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}"
+
+
+class Glm4MoeMoE(nn.Module):
+    """
+    A mixed expert module containing shared experts.
+    """
+
+    def __init__(self, config):
+        super().__init__()
+        self.config = config
+        self.experts = nn.ModuleList(
+            [
+                Glm4MoeMLP(config, intermediate_size=config.moe_intermediate_size)
+                for _ in range(config.n_routed_experts)
+            ]
+        )
+        self.gate = Glm4MoeTopkRouter(config)
+        self.shared_experts = Glm4MoeMLP(
+            config=config, intermediate_size=config.moe_intermediate_size * config.n_shared_experts
+        )
+
+    def moe(self, hidden_states: torch.Tensor, topk_indices: torch.Tensor, topk_weights: torch.Tensor):
+        r"""
+        CALL FOR CONTRIBUTION! I don't have time to optimise this right now, but expert weights need to be fused
+        to not have to do a loop here (deepseek has 256 experts soooo yeah).
+        """
+        final_hidden_states = torch.zeros_like(hidden_states, dtype=topk_weights.dtype)
+        expert_mask = torch.nn.functional.one_hot(topk_indices, num_classes=len(self.experts))
+        expert_mask = expert_mask.permute(2, 0, 1)
+
+        for expert_idx in range(len(self.experts)):
+            expert = self.experts[expert_idx]
+            mask = expert_mask[expert_idx]
+            token_indices, weight_indices = torch.where(mask)
+
+            if token_indices.numel() > 0:
+                expert_weights = topk_weights[token_indices, weight_indices]
+                expert_input = hidden_states[token_indices]
+                expert_output = expert(expert_input)
+                weighted_output = expert_output * expert_weights.unsqueeze(-1)
+                final_hidden_states.index_add_(0, token_indices, weighted_output)
+
+        # in original deepseek, the output of the experts are gathered once we leave this module
+        # thus the moe module is itelsf an IsolatedParallel module
+        # and all expert are "local" meaning we shard but we don't gather
+        return final_hidden_states.type(hidden_states.dtype)
+
+    def forward(self, hidden_states):
+        residuals = hidden_states
+        orig_shape = hidden_states.shape
+        topk_indices, topk_weights, router_logits = self.gate(hidden_states)
+        hidden_states = hidden_states.view(-1, hidden_states.shape[-1])
+        hidden_states = self.moe(hidden_states, topk_indices, topk_weights).view(*orig_shape)
+        hidden_states = hidden_states + self.shared_experts(residuals)
+        return hidden_states, router_logits
+
+
+class Glm4MoeDecoderLayer(GradientCheckpointingLayer):
+    def __init__(self, config: Glm4MoeConfig, layer_idx: int):
+        super().__init__()
+        self.hidden_size = config.hidden_size
+
+        self.self_attn = Glm4MoeAttention(config=config, layer_idx=layer_idx)
+
+        if layer_idx >= config.first_k_dense_replace:
+            self.mlp = Glm4MoeMoE(config)
+        else:
+            self.mlp = Glm4MoeMLP(config)
+
+        self.input_layernorm = Glm4MoeRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+        self.post_attention_layernorm = Glm4MoeRMSNorm(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[Cache] = None,
+        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[TransformersKwargs],
+    ) -> torch.Tensor:
+        residual = hidden_states
+        hidden_states = self.input_layernorm(hidden_states)
+        # Self Attention
+        hidden_states, _ = self.self_attn(
+            hidden_states=hidden_states,
+            attention_mask=attention_mask,
+            position_ids=position_ids,
+            past_key_value=past_key_value,
+            use_cache=use_cache,
+            cache_position=cache_position,
+            position_embeddings=position_embeddings,
+            **kwargs,
+        )
+        hidden_states = residual + hidden_states
+
+        # Fully Connected
+        residual = hidden_states
+        hidden_states = self.post_attention_layernorm(hidden_states)
+        mlp_output = self.mlp(hidden_states)
+        if len(mlp_output) == 2:
+            # If the MLP returns both hidden states and router logits
+            hidden_states, _ = mlp_output
+        else:
+            # If the MLP returns only hidden states
+            hidden_states = mlp_output
+        hidden_states = residual + hidden_states
+        return hidden_states
+
+
+@auto_docstring
+class Glm4MoePreTrainedModel(PreTrainedModel):
+    config: Glm4MoeConfig
+    base_model_prefix = "model"
+    supports_gradient_checkpointing = True
+    _no_split_modules = ["Glm4MoeDecoderLayer"]
+    _skip_keys_device_placement = ["past_key_values"]
+    _supports_flash_attn = True
+    _supports_sdpa = True
+    _supports_flex_attn = True
+    _can_compile_fullgraph = False
+    _supports_attention_backend = True
+    _can_record_outputs = {
+        "hidden_states": Glm4MoeDecoderLayer,
+        "attentions": Glm4MoeAttention,
+    }
+
+    def _init_weights(self, module):
+        super()._init_weights(module)
+        if isinstance(module, Glm4MoeTopkRouter):
+            module.weight.data.normal_(mean=0.0, std=self.config.initializer_range)
+
+
+class Glm4MoeRotaryEmbedding(nn.Module):
+    inv_freq: torch.Tensor  # fix linting for `register_buffer`
+
+    def __init__(self, config: Glm4MoeConfig, device=None):
+        super().__init__()
+        # BC: "rope_type" was originally "type"
+        if hasattr(config, "rope_scaling") and isinstance(config.rope_scaling, dict):
+            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)
+
+
+@auto_docstring
+class Glm4MoeModel(Glm4MoePreTrainedModel):
+    _keys_to_ignore_on_load_unexpected = [r"model\.layers\.92.*", r"model\.layers\.46.*"]
+
+    def __init__(self, config: Glm4MoeConfig):
+        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(
+            [Glm4MoeDecoderLayer(config, layer_idx) for layer_idx in range(config.num_hidden_layers)]
+        )
+        self.norm = Glm4MoeRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+        self.rotary_emb = Glm4MoeRotaryEmbedding(config=config)
+        self.gradient_checkpointing = False
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    @check_model_inputs
+    @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,
+        cache_position: Optional[torch.LongTensor] = None,
+        use_cache: Optional[bool] = None,
+        **kwargs: Unpack[TransformersKwargs],
+    ) -> BaseModelOutputWithPast:
+        if (input_ids is None) ^ (inputs_embeds is not None):
+            raise ValueError("You must specify exactly one of input_ids or inputs_embeds")
+
+        if inputs_embeds is None:
+            inputs_embeds: torch.Tensor = self.embed_tokens(input_ids)
+
+        if use_cache and past_key_values is None:
+            past_key_values = DynamicCache()
+
+        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.Tensor = 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)
+
+        causal_mask = create_causal_mask(
+            config=self.config,
+            input_embeds=inputs_embeds,
+            attention_mask=attention_mask,
+            cache_position=cache_position,
+            past_key_values=past_key_values,
+            position_ids=position_ids,
+        )
+
+        hidden_states = inputs_embeds
+        position_embeddings = self.rotary_emb(hidden_states, position_ids)
+
+        for decoder_layer in self.layers[: self.config.num_hidden_layers]:
+            hidden_states = decoder_layer(
+                hidden_states,
+                attention_mask=causal_mask,
+                position_ids=position_ids,
+                past_key_value=past_key_values,
+                cache_position=cache_position,
+                position_embeddings=position_embeddings,
+                **kwargs,
+            )
+
+        hidden_states = self.norm(hidden_states)
+        return BaseModelOutputWithPast(
+            last_hidden_state=hidden_states,
+            past_key_values=past_key_values,
+        )
+
+
+@auto_docstring
+class Glm4MoeForCausalLM(Glm4MoePreTrainedModel, GenerationMixin):
+    _tied_weights_keys = ["lm_head.weight"]
+    _tp_plan = {"lm_head": "colwise_rep"}
+    _pp_plan = {"lm_head": (["hidden_states"], ["logits"])}
+
+    def __init__(self, config):
+        super().__init__(config)
+        self.model = Glm4MoeModel(config)
+        self.vocab_size = config.vocab_size
+        self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    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[Cache] = None,
+        inputs_embeds: Optional[torch.FloatTensor] = None,
+        labels: Optional[torch.LongTensor] = None,
+        use_cache: Optional[bool] = None,
+        cache_position: Optional[torch.LongTensor] = None,
+        logits_to_keep: Union[int, torch.Tensor] = 0,
+        **kwargs: Unpack[TransformersKwargs],
+    ) -> CausalLMOutputWithPast:
+        r"""
+        Example:
+
+        ```python
+        >>> from transformers import AutoTokenizer, Glm4MoeForCausalLM
+
+        >>> model = Glm4MoeForCausalLM.from_pretrained("meta-glm4_moe/Glm4Moe-2-7b-hf")
+        >>> tokenizer = AutoTokenizer.from_pretrained("meta-glm4_moe/Glm4Moe-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."
+        ```"""
+        outputs: BaseModelOutputWithPast = 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,
+            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=logits, labels=labels, vocab_size=self.config.vocab_size, **kwargs)
+
+        return CausalLMOutputWithPast(
+            loss=loss,
+            logits=logits,
+            past_key_values=outputs.past_key_values,
+            hidden_states=outputs.hidden_states,
+            attentions=outputs.attentions,
+        )
+
+
+__all__ = ["Glm4MoePreTrainedModel", "Glm4MoeModel", "Glm4MoeForCausalLM"]

reap_args.yaml

@@ -0,0 +1,76 @@
+cluster_args:
+  cluster_description: null
+  cluster_method: agglomerative
+  compression_ratio: 0.5
+  expert_sim: ttm
+  frequency_penalty: true
+  linkage_method: average
+  max_cluster_size: null
+  multi_layer: null
+  num_clusters: null
+  singleton_outlier_experts: false
+  singleton_super_experts: false
+  softmax_temperature: null
+ds_args:
+  dataset_config_name: all
+  dataset_name: theblackcat102/evol-codealpaca-v1
+  dataset_test_split: test
+  shuffle: true
+  split: train
+eval_args:
+  evalplus_tasks:
+  - mbpp
+  - humaneval
+  greedy: true
+  lm_eval_tasks:
+  - winogrande
+  - arc_challenge
+  - arc_easy
+  - boolq
+  - hellaswag
+  - mmlu
+  - openbookqa
+  - rte
+  min_p: 0.0
+  parallel_tasks: 32
+  results_dir: null
+  run_evalplus: true
+  run_livecodebench: true
+  run_lm_eval: true
+  run_math: false
+  run_wildbench: false
+  server_log_file_name: pruning-cli-0.log
+  temperature: 0.7
+  top_k: 20
+  top_p: 0.8
+  use_server: true
+  vllm_port: 8000
+model_args:
+  model_name: zai-org/GLM-4.6
+  num_experts_per_tok_override: null
+obs_args:
+  distance_measure: cosine
+  model_max_length: 2048
+  output_file_name: observations_10_cosine-seed_42.pt
+  overwrite_observations: false
+  record_pruning_metrics_only: true
+  renormalize_router_weights: false
+  return_vllm_tokens_prompt: false
+  samples_per_category: 10
+  select_only_categories: null
+  split_by_category: false
+  truncate: false
+prune_args:
+  n_experts_to_prune: null
+  overwrite_pruned_model: false
+  perserve_outliers: false
+  perserve_super_experts: false
+  prune_method: reap
+reap_args:
+  debug: false
+  do_eval: false
+  plot_clusters: true
+  profile: false
+  run_observer_only: false
+  seed: 42
+  smoke_test: true

special_tokens_map.json

@@ -0,0 +1,40 @@
+{
+  "additional_special_tokens": [
+    "<|endoftext|>",
+    "[MASK]",
+    "[gMASK]",
+    "[sMASK]",
+    "<sop>",
+    "<eop>",
+    "<|system|>",
+    "<|user|>",
+    "<|assistant|>",
+    "<|observation|>",
+    "<|begin_of_image|>",
+    "<|end_of_image|>",
+    "<|begin_of_video|>",
+    "<|end_of_video|>",
+    "<|begin_of_audio|>",
+    "<|end_of_audio|>",
+    "<|begin_of_transcription|>",
+    "<|end_of_transcription|>",
+    "<|code_prefix|>",
+    "<|code_middle|>",
+    "<|code_suffix|>",
+    "/nothink"
+  ],
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bda8e2146c3bb7b7e0fc96dcc4f0aeff041c6c27952e3ace0665663ebff346ba
+size 19970700