Upload folder using huggingface_hub

.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

config.json

@@ -0,0 +1,175 @@
+{
+    "_commit_hash": null,
+    "architectures": [
+        "POINTSGUIModel"
+    ],
+    "auto_map": {
+        "AutoConfig": "configuration_points_gui.POINTSGUIConfig",
+        "AutoModelForCausalLM": "modeling_points_gui.POINTSGUIModel",
+        "AutoModel": "modeling_points_gui.POINTSGUIModel"
+    },
+    "llm_config": {
+        "_attn_implementation_autoset": true,
+        "_name_or_path": "",
+        "add_cross_attention": false,
+        "architectures": null,
+        "attention_bias": false,
+        "attention_dropout": 0.0,
+        "bad_words_ids": null,
+        "begin_suppress_tokens": null,
+        "bos_token_id": null,
+        "chunk_size_feed_forward": 0,
+        "cross_attention_hidden_size": null,
+        "decoder_start_token_id": null,
+        "diversity_penalty": 0.0,
+        "do_sample": false,
+        "early_stopping": false,
+        "encoder_no_repeat_ngram_size": 0,
+        "eos_token_id": null,
+        "exponential_decay_length_penalty": null,
+        "finetuning_task": null,
+        "forced_bos_token_id": null,
+        "forced_eos_token_id": null,
+        "head_dim": 128,
+        "hidden_act": "silu",
+        "hidden_size": 4096,
+        "id2label": {
+            "0": "LABEL_0",
+            "1": "LABEL_1"
+        },
+        "initializer_range": 0.02,
+        "intermediate_size": 12288,
+        "is_decoder": false,
+        "is_encoder_decoder": false,
+        "label2id": {
+            "LABEL_0": 0,
+            "LABEL_1": 1
+        },
+        "length_penalty": 1.0,
+        "max_length": 20,
+        "max_position_embeddings": 32768,
+        "max_window_layers": 28,
+        "min_length": 0,
+        "model_type": "qwen3",
+        "no_repeat_ngram_size": 0,
+        "num_attention_heads": 32,
+        "num_beam_groups": 1,
+        "num_beams": 1,
+        "num_hidden_layers": 36,
+        "num_key_value_heads": 8,
+        "num_return_sequences": 1,
+        "output_attentions": false,
+        "output_hidden_states": false,
+        "output_scores": false,
+        "pad_token_id": null,
+        "prefix": null,
+        "problem_type": null,
+        "pruned_heads": {},
+        "remove_invalid_values": false,
+        "repetition_penalty": 1.0,
+        "return_dict": true,
+        "return_dict_in_generate": false,
+        "rms_norm_eps": 1e-06,
+        "rope_scaling": null,
+        "rope_theta": 1000000.0,
+        "sep_token_id": null,
+        "sliding_window": 4096,
+        "suppress_tokens": null,
+        "task_specific_params": null,
+        "temperature": 1.0,
+        "tf_legacy_loss": false,
+        "tie_encoder_decoder": false,
+        "tie_word_embeddings": false,
+        "tokenizer_class": null,
+        "top_k": 50,
+        "top_p": 1.0,
+        "torch_dtype": null,
+        "torchscript": false,
+        "transformers_version": "4.51.0",
+        "typical_p": 1.0,
+        "use_bfloat16": false,
+        "use_cache": true,
+        "use_sliding_window": false,
+        "vocab_size": 151936
+    },
+    "torch_dtype": "bfloat16",
+    "transformers_version": null,
+    "vision_config": {
+        "_attn_implementation_autoset": false,
+        "_name_or_path": "",
+        "add_cross_attention": false,
+        "architectures": null,
+        "bad_words_ids": null,
+        "begin_suppress_tokens": null,
+        "bos_token_id": null,
+        "chunk_size_feed_forward": 0,
+        "cross_attention_hidden_size": null,
+        "decoder_start_token_id": null,
+        "depth": 32,
+        "diversity_penalty": 0.0,
+        "do_sample": false,
+        "early_stopping": false,
+        "embed_dim": 1280,
+        "encoder_no_repeat_ngram_size": 0,
+        "eos_token_id": null,
+        "exponential_decay_length_penalty": null,
+        "finetuning_task": null,
+        "forced_bos_token_id": null,
+        "forced_eos_token_id": null,
+        "hidden_act": "quick_gelu",
+        "hidden_size": 3584,
+        "id2label": {
+            "0": "LABEL_0",
+            "1": "LABEL_1"
+        },
+        "in_channels": 3,
+        "in_chans": 3,
+        "is_decoder": false,
+        "is_encoder_decoder": false,
+        "label2id": {
+            "LABEL_0": 0,
+            "LABEL_1": 1
+        },
+        "length_penalty": 1.0,
+        "max_length": 20,
+        "min_length": 0,
+        "mlp_ratio": 4,
+        "model_type": "qwen2_vl",
+        "no_repeat_ngram_size": 0,
+        "num_beam_groups": 1,
+        "num_beams": 1,
+        "num_heads": 16,
+        "num_return_sequences": 1,
+        "output_attentions": false,
+        "output_hidden_states": false,
+        "output_scores": false,
+        "pad_token_id": null,
+        "patch_size": 14,
+        "prefix": null,
+        "problem_type": null,
+        "pruned_heads": {},
+        "remove_invalid_values": false,
+        "repetition_penalty": 1.0,
+        "return_dict": true,
+        "return_dict_in_generate": false,
+        "sep_token_id": null,
+        "spatial_merge_size": 2,
+        "spatial_patch_size": 14,
+        "suppress_tokens": null,
+        "task_specific_params": null,
+        "temperature": 1.0,
+        "temporal_patch_size": 2,
+        "tf_legacy_loss": false,
+        "tie_encoder_decoder": false,
+        "tie_word_embeddings": true,
+        "tokenizer_class": null,
+        "top_k": 50,
+        "top_p": 1.0,
+        "torch_dtype": "bfloat16",
+        "torchscript": false,
+        "transformers_version": "4.51.0",
+        "typical_p": 1.0,
+        "use_bfloat16": false
+    },
+    "image_token_id": 151655
+}

configuration_points_gui.py

@@ -0,0 +1,66 @@
+from typing import Any
+
+from transformers import PretrainedConfig, Qwen3Config
+
+try:
+    from transformers.models.qwen2_vl.configuration_qwen2_vl import Qwen2VLVisionConfig
+except ImportError:
+    print('Please upgrade transformers to version 4.46.3 or higher')
+
+
+class POINTSGUIConfig(PretrainedConfig):
+    model_type = "points_gui"
+    is_composition = True
+    """Configuration class for `POINTSGUI`."""
+
+    def __init__(self,
+                 **kwargs) -> None:
+        super().__init__(**kwargs)
+        if not kwargs:
+            return
+        vision_config = kwargs.pop("vision_config", None)
+        llm_config = kwargs.pop("llm_config", None)
+        if isinstance(vision_config, dict):
+            self.vision_config = Qwen2VLVisionConfig(**vision_config)
+        else:
+            self.vision_config = vision_config
+        if isinstance(llm_config, dict):
+            self.llm_config = Qwen3Config(**llm_config)
+        else:
+            self.llm_config = llm_config
+
+        self.vocab_size = llm_config["vocab_size"]
+        self.max_position_embeddings = llm_config["max_position_embeddings"]
+        self.hidden_size = llm_config["hidden_size"]
+        self.intermediate_size = llm_config["intermediate_size"]
+        self.num_hidden_layers = llm_config["num_hidden_layers"]
+        self.num_attention_heads = llm_config["num_attention_heads"]
+        self.use_sliding_window = llm_config["use_sliding_window"]
+        self.sliding_window = llm_config["sliding_window"]  # we check `use_sliding_window` in the modeling code
+        self.max_window_layers = llm_config["max_window_layers"]
+
+        # for backward compatibility
+        if llm_config["num_key_value_heads"] is None:
+            llm_config["num_key_value_heads"] = llm_config["num_attention_heads"]
+
+        self.num_key_value_heads = llm_config["num_key_value_heads"]
+        self.head_dim = llm_config["head_dim"]
+        self.hidden_act = llm_config["hidden_act"]
+        self.initializer_range = llm_config["initializer_range"]
+        self.rms_norm_eps = llm_config["rms_norm_eps"]
+        self.use_cache = llm_config["use_cache"]
+        self.rope_theta = llm_config["rope_theta"]
+        self.rope_scaling = llm_config["rope_scaling"]
+        self.attention_bias = llm_config["attention_bias"]
+        self.attention_dropout = llm_config["attention_dropout"]
+        # Validate the correctness of rotary position embeddings parameters
+        # BC: if there is a 'type' field, move it to 'rope_type'.
+        if self.rope_scaling is not None and "type" in self.rope_scaling:
+            if self.rope_scaling["type"] == "mrope":
+                self.rope_scaling["type"] = "default"
+            self.rope_scaling["rope_type"] = self.rope_scaling["type"]
+
+        super().__init__(
+            tie_word_embeddings=llm_config["tie_word_embeddings"],
+            **kwargs,
+        )

generation_config.json

@@ -0,0 +1,10 @@
+{
+  "_from_model_config": true,
+  "transformers_version": "4.51.0",
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "bos_token_id": 151643,
+  "pad_token_id": 151643
+}

mergekit_config.yml

@@ -0,0 +1,17 @@
+models:
+  - model: /mnt/cephfs/bensenliu/wfs/weights/rl/202601/20260113e3-hf
+    parameters:
+      weight: 1.0
+  - model: /mnt/cephfs/bensenliu/wfs/weights/rl/202601/20260119e3-hf
+    parameters:
+      weight: 1.0
+  - model: /mnt/cephfs/bensenliu/wfs/weights/rl/202601/20260121e1-hf
+    parameters:
+      weight: 1.0
+  - model: /mnt/cephfs/bensenliu/wfs/weights/rl/202601/20260124e1-hf
+    parameters:
+      weight: 1.0
+merge_method: linear
+parameters:
+  normalize: true
+dtype: bfloat16

model-00001-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ade58cd7348d41db49c5bb9eb6497bb824c6299593093011b6f60aaa4917de2e
+size 4905356464

model-00002-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:88e78b868119ae33a26a8dc3592ec0532d5ddfdd83e24a4cc717f6b03b8a7810
+size 4915960936

model-00003-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:317748ffe26ba4726fb3e2a826f3c106916e3c2270b488ac02e9b19d966c7650
+size 4974672744

model-00004-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dfa53cbe4d3279687d982a458b4414d53202ed2d88f84945b2590a6979cb505b
+size 3031486688

model.safetensors.index.json

@@ -0,0 +1,803 @@
+{
+  "metadata": {
+    "total_size": 17827384320
+  },
+  "weight_map": {
+    "llm.lm_head.weight": "model-00004-of-00004.safetensors",
+    "llm.model.embed_tokens.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.0.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.0.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.0.self_attn.k_norm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.0.self_attn.q_norm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.1.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.1.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.1.self_attn.k_norm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.1.self_attn.q_norm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.10.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.10.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.10.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.10.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.11.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.11.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.11.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.11.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.12.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.12.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.12.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.12.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.13.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.13.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.13.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.13.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.14.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.14.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.14.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.14.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.15.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.15.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.15.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.15.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.16.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.16.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.16.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.16.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.16.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.16.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.16.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.17.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.17.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.17.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.17.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.17.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.17.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.17.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.17.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.18.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.18.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.18.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.18.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.18.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.18.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.18.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.18.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.18.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.19.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.19.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.19.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.19.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.19.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.19.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.19.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.19.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.19.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.19.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.19.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.2.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.2.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.2.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.2.self_attn.k_norm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.2.self_attn.q_norm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.20.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.20.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.20.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.20.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.20.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.20.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.20.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.20.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.20.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.20.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.20.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.21.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.21.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.21.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.21.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.21.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.21.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.21.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.21.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.21.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.21.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.21.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.22.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.22.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.22.self_attn.k_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.22.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.22.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.22.self_attn.q_norm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.22.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.22.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.23.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.23.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.23.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.23.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.24.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.24.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.24.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.24.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.25.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.25.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.25.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.25.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.26.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.26.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.26.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.26.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.26.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.27.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.27.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.27.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.27.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.27.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.28.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.28.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.28.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.28.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.28.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.28.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.28.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.28.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.28.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.28.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.28.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.29.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.29.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.29.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.29.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.29.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.29.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.29.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.29.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.29.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.29.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.29.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.3.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.3.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.3.self_attn.k_norm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.3.self_attn.q_norm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.30.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.30.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.30.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.30.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.30.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.30.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.30.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.30.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.30.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.30.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.30.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.31.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.31.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.31.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.31.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.31.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.31.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.31.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.31.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.31.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.31.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.31.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.32.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.32.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.32.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.32.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.32.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.32.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.32.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.32.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.32.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.32.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.32.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.33.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.33.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.33.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.33.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.33.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.33.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.33.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.33.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.33.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.33.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.33.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.34.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.34.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.34.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.34.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.34.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.34.self_attn.k_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.34.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.34.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.34.self_attn.q_norm.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.34.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.34.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.35.input_layernorm.weight": "model-00004-of-00004.safetensors",
+    "llm.model.layers.35.mlp.down_proj.weight": "model-00004-of-00004.safetensors",
+    "llm.model.layers.35.mlp.gate_proj.weight": "model-00004-of-00004.safetensors",
+    "llm.model.layers.35.mlp.up_proj.weight": "model-00004-of-00004.safetensors",
+    "llm.model.layers.35.post_attention_layernorm.weight": "model-00004-of-00004.safetensors",
+    "llm.model.layers.35.self_attn.k_norm.weight": "model-00004-of-00004.safetensors",
+    "llm.model.layers.35.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.35.self_attn.o_proj.weight": "model-00004-of-00004.safetensors",
+    "llm.model.layers.35.self_attn.q_norm.weight": "model-00004-of-00004.safetensors",
+    "llm.model.layers.35.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.35.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "llm.model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.4.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.4.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.4.self_attn.k_norm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.4.self_attn.q_norm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.5.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.5.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.5.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.5.self_attn.k_norm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.5.self_attn.q_norm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.6.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.6.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.6.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.6.self_attn.k_norm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.6.self_attn.q_norm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.7.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.7.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.7.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.7.self_attn.k_norm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.7.self_attn.q_norm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.8.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.8.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.8.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.8.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.8.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.8.self_attn.k_norm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.8.self_attn.q_norm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.9.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.9.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.9.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.9.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "llm.model.layers.9.self_attn.k_norm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.9.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.9.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.9.self_attn.q_norm.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.9.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.layers.9.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "llm.model.norm.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.0.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.0.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.0.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.0.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.0.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.0.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.0.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.0.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.0.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.0.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.0.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.0.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.1.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.1.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.1.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.1.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.1.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.1.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.1.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.1.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.1.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.1.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.1.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.1.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.10.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.10.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.10.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.10.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.10.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.10.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.10.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.10.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.10.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.10.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.10.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.10.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.11.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.11.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.11.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.11.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.11.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.11.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.11.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.11.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.11.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.11.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.11.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.11.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.12.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.12.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.12.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.12.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.12.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.12.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.12.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.12.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.12.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.12.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.12.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.12.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.13.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.13.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.13.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.13.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.13.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.13.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.13.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.13.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.13.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.13.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.13.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.13.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.14.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.14.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.14.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.14.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.14.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.14.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.14.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.14.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.14.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.14.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.14.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.14.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.15.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.15.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.15.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.15.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.15.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.15.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.15.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.15.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.15.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.15.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.15.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.15.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.16.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.16.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.16.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.16.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.16.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.16.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.16.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.16.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.16.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.16.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.16.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.16.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.17.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.17.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.17.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.17.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.17.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.17.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.17.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.17.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.17.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.17.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.17.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.17.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.18.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.18.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.18.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.18.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.18.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.18.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.18.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.18.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.18.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.18.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.18.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.18.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.19.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.19.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.19.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.19.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.19.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.19.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.19.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.19.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.19.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.19.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.19.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.19.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.2.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.2.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.2.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.2.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.2.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.2.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.2.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.2.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.2.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.2.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.2.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.2.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.20.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.20.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.20.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.20.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.20.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.20.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.20.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.20.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.20.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.20.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.20.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.20.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.21.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.21.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.21.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.21.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.21.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.21.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.21.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.21.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.21.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.21.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.21.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.21.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.22.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.22.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.22.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.22.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.22.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.22.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.22.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.22.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.22.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.22.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.22.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.22.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.23.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.23.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.23.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.23.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.23.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.23.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.23.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.23.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.23.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.23.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.23.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.23.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.24.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.24.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.24.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.24.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.24.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.24.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.24.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.24.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.24.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.24.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.24.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.24.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.25.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.25.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.25.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.25.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.25.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.25.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.25.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.25.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.25.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.25.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.25.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.25.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.26.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.26.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.26.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.26.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.26.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.26.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.26.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.26.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.26.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.26.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.26.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.26.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.27.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.27.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.27.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.27.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.27.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.27.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.27.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.27.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.27.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.27.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.27.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.27.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.28.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.28.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.28.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.28.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.28.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.28.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.28.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.28.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.28.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.28.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.28.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.28.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.29.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.29.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.29.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.29.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.29.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.29.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.29.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.29.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.29.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.29.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.29.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.29.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.3.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.3.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.3.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.3.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.3.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.3.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.3.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.3.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.3.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.3.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.3.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.3.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.30.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.30.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.30.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.30.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.30.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.30.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.30.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.30.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.30.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.30.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.30.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.30.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.31.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.31.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.31.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.31.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.31.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.31.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.31.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.31.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.31.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.31.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.31.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.31.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.4.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.4.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.4.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.4.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.4.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.4.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.4.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.4.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.4.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.4.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.4.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.4.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.5.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.5.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.5.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.5.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.5.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.5.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.5.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.5.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.5.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.5.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.5.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.5.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.6.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.6.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.6.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.6.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.6.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.6.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.6.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.6.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.6.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.6.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.6.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.6.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.7.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.7.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.7.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.7.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.7.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.7.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.7.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.7.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.7.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.7.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.7.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.7.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.8.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.8.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.8.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.8.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.8.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.8.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.8.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.8.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.8.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.8.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.8.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.8.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.9.attn.proj.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.9.attn.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.9.attn.qkv.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.9.attn.qkv.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.9.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.9.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.9.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.9.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.9.norm1.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.9.norm1.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.9.norm2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.blocks.9.norm2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.merger.ln_q.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.merger.ln_q.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.merger.mlp.0.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.merger.mlp.0.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.merger.mlp.2.bias": "model-00004-of-00004.safetensors",
+    "vision_encoder.merger.mlp.2.weight": "model-00004-of-00004.safetensors",
+    "vision_encoder.patch_embed.proj.weight": "model-00004-of-00004.safetensors",
+    "vision_projector.ln_q.bias": "model-00004-of-00004.safetensors",
+    "vision_projector.ln_q.weight": "model-00004-of-00004.safetensors",
+    "vision_projector.mlp.0.bias": "model-00004-of-00004.safetensors",
+    "vision_projector.mlp.0.weight": "model-00004-of-00004.safetensors",
+    "vision_projector.mlp.2.bias": "model-00004-of-00004.safetensors",
+    "vision_projector.mlp.2.weight": "model-00004-of-00004.safetensors"
+  }
+}

modeling_points_gui.py

@@ -0,0 +1,321 @@
+import math
+from typing import List, Optional, Tuple, Union, Any
+
+import numpy as np
+import torch
+from PIL import Image
+from transformers import (
+    GenerationMixin,
+    PreTrainedModel,
+    PreTrainedTokenizer
+)
+
+try:
+    from transformers import Qwen3ForCausalLM
+except ImportError:
+    print('Please upgrade transformers to version 4.51.0 or higher')
+
+try:
+    from transformers.models.qwen2_vl.image_processing_qwen2_vl import (  # noqa
+        Qwen2VLImageProcessor,
+    )
+    from transformers.models.qwen2_vl.modeling_qwen2_vl import PatchMerger
+except ImportError:
+    print('Please upgrade transformers to version 4.46.3 or higher')
+
+from .configuration_points_gui import POINTSGUIConfig
+
+try:
+    from wepoints.models import Qwen2VisionTransformerForNavitPOINTS
+except ImportError:
+    print('Please install WePOINTS, and refer to https://github.com/WePOINTS/WePOINTS')
+from transformers.models.qwen2_vl.modeling_qwen2_vl import Qwen2VLCausalLMOutputWithPast
+
+
+class POINTSGUIModel(PreTrainedModel, GenerationMixin):
+    config_class = POINTSGUIConfig
+    _no_split_modules = []
+    _supports_flash_attn_2 = True
+    supports_gradient_checkpointing = True
+    """Chat model for POINTSv1.5.
+    
+    Args:
+        config (POINTSChatConfigV15): The model config.
+    """
+
+    def __init__(self, config: POINTSGUIConfig, **kwargs) -> None:
+        super().__init__(config)
+        config.llm_config._attn_implementation = "flash_attention_2"
+        config._attn_implementation_autoset = False
+        self.llm = Qwen3ForCausalLM(config.llm_config)
+        self.vision_encoder = Qwen2VisionTransformerForNavitPOINTS._from_config(  # noqa
+            config.vision_config, attn_implementation="flash_attention_2"
+        )
+        self.vision_projector = PatchMerger(config.llm_config.hidden_size,
+                                            context_dim=1280).to(torch.bfloat16)
+        
+    def process_images(self, images: torch.Tensor, 
+                       image_grid_thws: List[list]) -> torch.Tensor:
+        """Obtain image features from the vision encoder.
+        
+        Args:
+            images (torch.Tensor): The input images.
+            image_grid_thws (List[list]): The grid thresholds for the images.
+
+        Returns:
+            torch.Tensor: The image features.
+        """
+        image_features = self.vision_encoder(images, grid_thw=image_grid_thws)
+        image_features = self.vision_projector(image_features)
+        return image_features
+    
+    def construct_prompt(self, messages: List[dict],
+                         image_processor: Qwen2VLImageProcessor) -> Tuple[str, List[Image.Image], List[list]]: # noqa
+        """Construct the prompt for the chat model.
+
+        Args:
+            messages (List[dict]): The input messages.
+        
+        Returns:
+            Tuple[str, List[Image.Image], List[list]]: 
+                The prompt, images, and image grid shape.
+        """
+        images = []
+        image_grid_thws = []
+        reconstructed_messages = []
+        for message in messages:
+            role = message['role']
+            content_from_role = ''
+            for item in message['content']:
+                if item['type'] == 'text':
+                    content_from_role += item['text']
+                elif item['type'] == 'image':
+                    image_path = item['image']
+                    max_pixels = item['max_pixels'] if 'max_pixels' in item else None
+                    image = Image.open(image_path).convert('RGB')
+                    if max_pixels is not None:
+                        # obtain image size
+                        width, height = image.size
+                        cur_image_pixels = width * height
+                        if cur_image_pixels > max_pixels:
+                            beta = math.sqrt((height * width) / max_pixels)
+                            new_width = math.floor(width / beta)
+                            new_height = math.floor(height / beta)
+                            image = image.resize((new_width, new_height))
+                    image_data = image_processor(images=image)
+                    pixel_values = image_data['pixel_values']
+                    image_grid_thw = image_data['image_grid_thw']
+                    images.extend(pixel_values)
+                    image_grid_thws.append(image_grid_thw)
+                    seq_len = int(image_grid_thw[0][1] * image_grid_thw[0][2] / 4) # noqa
+                    content_from_role += '<|vision_start|>' + '<|image_pad|>' * seq_len + '<|vision_end|>' + '\n' # noqa
+            reconstructed_messages.append({
+                'role': role,
+                'content': content_from_role
+            })
+        prompt = self.apply_chat_template(reconstructed_messages)
+        return prompt, images, image_grid_thws
+    
+    def apply_chat_template(self, messages: List[dict]) -> str:
+        """Apply the chat template to the input messages.
+
+        Args:
+            messages (List[dict]): The input messages.
+        
+        Returns:
+            str: The prompt.
+        """
+        role_prefix_mapping = {
+            'user': '<|im_start|>user\n',
+            'assistant': '<|im_start|>assistant\n',
+            'system': '<|im_start|>system\n'
+        }
+        role = 'user'
+        prompt = ''
+        for message in messages:
+            role = message['role']
+            content = message['content']
+            prompt += role_prefix_mapping[role] + content + '<|im_end|>\n'
+        if role == 'user':
+            prompt += '<|im_start|>assistant\n'
+        return prompt
+
+    @torch.no_grad()
+    def chat(self, 
+             messages: List[dict],
+             tokenizer: PreTrainedTokenizer,
+             image_processor: object,
+             generation_config: dict = None) -> str:
+        """Generate a response to the input prompt.
+
+        Args:
+            messages (List[dict]): The input messages.
+            tokenizer (PreTrainedTokenizer): The tokenizer to use.
+            image_processor (object): The image processor to use.
+            generation_config (dict, optional): The generation config. 
+                Defaults to None.
+        Returns:
+            str: The generated response.
+        """
+        prompt, images, image_grid_thws = self.construct_prompt(
+            messages, image_processor
+        )
+        images = np.array(images)
+        images = torch.from_numpy(images).to(self.vision_encoder.device).to(self.vision_encoder.dtype) # noqa
+        image_grid_thws = np.concatenate(image_grid_thws, axis=0)
+        image_grid_thws = (
+            torch.from_numpy(image_grid_thws)
+            .cuda()
+            .long()
+        )
+        image_features = self.vision_encoder(images, grid_thw=image_grid_thws)
+        
+        image_features = self.vision_projector(image_features)
+        model_inputs = tokenizer(prompt, return_tensors='pt')
+        input_ids = model_inputs['input_ids'].to(self.device)
+        attention_mask = model_inputs['attention_mask'].to(self.device)
+        # stop token
+        eos_token_id = tokenizer.convert_tokens_to_ids("<|im_end|>")
+        # image token
+        image_token_id = tokenizer.convert_tokens_to_ids("<|image_pad|>")
+        generation_config.update(
+            {
+                'eos_token_id': eos_token_id,
+            }
+        )
+        outputs = self.generate(
+            input_ids=input_ids,
+            image_grid_thws=image_grid_thws,
+            attention_mask=attention_mask,
+            image_features=[image_features],
+            image_token_id=image_token_id,
+            **generation_config
+        )
+        response = tokenizer.batch_decode(outputs, skip_special_tokens=True)[0]
+        return response
+      
+    def _split_input_ids(self, input_ids, special_token):
+        special_pos = input_ids == special_token
+        pos = (special_pos[:-1] != special_pos[1:]).nonzero() + 1
+        if pos.shape[0] % 2 != 0:
+            pos = torch.cat([torch.tensor([[0]]).to(pos.device), pos])
+        pos = pos.reshape(-1, 2).tolist()
+        return pos
+
+    def generate(self,
+                 input_ids: torch.LongTensor,
+                 image_grid_thws: torch.LongTensor,
+                 attention_mask: torch.LongTensor,
+                 image_features: List[torch.Tensor],
+                 image_token_id: int,
+                 generation_config: Optional[dict] = None,
+                 output_hidden_states: Optional[bool] = None,
+                 **generate_kwargs) -> torch.LongTensor:
+        input_embeddings = self.llm.model.embed_tokens(input_ids)
+        batch_size = input_ids.shape[0]
+        assert len(image_features) == batch_size
+        for i in range(batch_size):
+            pos = self._split_input_ids(input_ids[i], image_token_id)
+            assert len(pos) == len(image_grid_thws)
+            image_pos = [
+                int(image_grid_thw[1] * image_grid_thw[2] / 4)
+                for image_grid_thw in image_grid_thws
+            ]
+            image_pos.insert(0, 0)
+            image_pos = np.cumsum(image_pos)
+            for j, (start, end) in enumerate(pos):
+                input_embeddings[i, start:end] = \
+                    image_features[i][image_pos[j]:image_pos[j+1]]
+        outputs = self.llm.generate(
+            inputs_embeds=input_embeddings,
+            attention_mask=attention_mask,
+            generation_config=generation_config,
+            output_hidden_states=output_hidden_states,
+            use_cache=True,
+            **generate_kwargs
+        )
+        return outputs
+
+    def get_image_features(self, pixel_values: torch.FloatTensor, image_grid_thw: Optional[torch.LongTensor] = None):
+        """
+        Encodes images into continuous embeddings that can be forwarded to the language model.
+
+        Args:
+            pixel_values (`torch.FloatTensor` of shape `(batch_size, num_channels, image_size, image_size)`):
+                The tensors corresponding to the input images.
+            image_grid_thw (`torch.LongTensor` of shape `(num_images, 3)`, *optional*):
+                The temporal, height and width of feature shape of each image in LLM.
+        """
+        pixel_values = pixel_values.type(self.visual.dtype)
+        image_embeds = self.visual(pixel_values, grid_thw=image_grid_thw)
+        split_sizes = (image_grid_thw.prod(-1) // self.visual.spatial_merge_size**2).tolist()
+        image_embeds = torch.split(image_embeds, split_sizes)
+        return image_embeds
+    
+    def forward(
+        self,
+        input_ids: 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,
+        return_dict: Optional[bool] = None,
+        pixel_values: Optional[torch.Tensor] = None,
+        pixel_values_videos: Optional[torch.FloatTensor] = None,
+        image_grid_thw: Optional[torch.LongTensor] = None,
+        video_grid_thw: Optional[torch.LongTensor] = None,
+        rope_deltas: Optional[torch.LongTensor] = None,
+        cache_position: Optional[torch.LongTensor] = None,
+        **kwargs: Optional[Any],
+    ) -> Union[Tuple, Qwen2VLCausalLMOutputWithPast]:
+
+        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
+        output_hidden_states = (
+            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
+        )
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+        if inputs_embeds is None:
+            inputs_embeds = self.llm.get_input_embeddings()(input_ids)
+            if pixel_values is not None:
+                image_embeds = self.process_images(pixel_values, image_grid_thw)
+                n_image_tokens = (input_ids == self.config.image_token_id).sum().item()
+                n_image_features = image_embeds.shape[0]
+                if n_image_tokens != n_image_features:
+                    raise ValueError(
+                        f"Image features and image tokens do not match: tokens: {n_image_tokens}, features {n_image_features}"
+                    )
+                image_mask = (
+                    (input_ids == self.config.image_token_id)
+                    .unsqueeze(-1)
+                    .expand_as(inputs_embeds)
+                    .to(inputs_embeds.device)
+                )
+                image_embeds = image_embeds.to(inputs_embeds.device, inputs_embeds.dtype)
+                inputs_embeds = inputs_embeds.masked_scatter(image_mask, image_embeds)
+        
+        
+        outputs = self.llm.forward(
+            input_ids=None,
+            inputs_embeds=inputs_embeds,
+            attention_mask=attention_mask,
+            output_hidden_states=output_hidden_states,
+            position_ids=position_ids,
+            past_key_values=past_key_values,
+            labels=labels,
+            use_cache=True,
+            output_attentions=output_attentions,
+            cache_position=cache_position,
+            **kwargs,
+        )
+
+        return Qwen2VLCausalLMOutputWithPast(
+            loss=outputs.loss,
+            logits=outputs.logits,
+            past_key_values=outputs.past_key_values,
+            hidden_states=outputs.hidden_states,
+            attentions=outputs.attentions
+        )

preprocessor_config.json

@@ -0,0 +1,31 @@
+{
+    "do_convert_rgb": true,
+    "do_normalize": true,
+    "do_rescale": true,
+    "do_resize": true,
+    "image_mean": [
+        0.48145466,
+        0.4578275,
+        0.40821073
+    ],
+    "image_processor_type": "Qwen2VLImageProcessor",
+    "image_std": [
+        0.26862954,
+        0.26130258,
+        0.27577711
+    ],
+    "max_pixels": 4233600,
+    "merge_size": 2,
+    "min_pixels": 3136,
+    "patch_size": 14,
+    "resample": 3,
+    "rescale_factor": 0.00392156862745098,
+    "size": {
+        "max_pixels": 4233600,
+        "min_pixels": 3136,
+        "longest_edge": 4233600,
+        "shortest_edge": 3136
+    },
+    "temporal_patch_size": 2,
+    "processor_class": "Qwen2VLProcessor"
+}

special_tokens_map.json

@@ -0,0 +1,31 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "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:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4
+size 11422654

tokenizer_config.json

@@ -0,0 +1,239 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151665": {
+      "content": "<tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151666": {
+      "content": "</tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151667": {
+      "content": "<think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151668": {
+      "content": "</think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0].role == 'system' %}\n        {{- messages[0].content + '\\n\\n' }}\n    {%- endif %}\n    {{- \"# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0].role == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0].content + '<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}\n{%- for message in messages[::-1] %}\n    {%- set index = (messages|length - 1) - loop.index0 %}\n    {%- if ns.multi_step_tool and message.role == \"user\" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}\n        {%- set ns.multi_step_tool = false %}\n        {%- set ns.last_query_index = index %}\n    {%- endif %}\n{%- endfor %}\n{%- for message in messages %}\n    {%- if message.content is string %}\n        {%- set content = message.content %}\n    {%- else %}\n        {%- set content = '' %}\n    {%- endif %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) %}\n        {{- '<|im_start|>' + message.role + '\\n' + content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {%- set reasoning_content = '' %}\n        {%- if message.reasoning_content is string %}\n            {%- set reasoning_content = message.reasoning_content %}\n        {%- else %}\n            {%- if '</think>' in content %}\n                {%- set reasoning_content = content.split('</think>')[0].rstrip('\\n').split('<think>')[-1].lstrip('\\n') %}\n                {%- set content = content.split('</think>')[-1].lstrip('\\n') %}\n            {%- endif %}\n        {%- endif %}\n        {%- if loop.index0 > ns.last_query_index %}\n            {%- if loop.last or (not loop.last and reasoning_content) %}\n                {{- '<|im_start|>' + message.role + '\\n<think>\\n' + reasoning_content.strip('\\n') + '\\n</think>\\n\\n' + content.lstrip('\\n') }}\n            {%- else %}\n                {{- '<|im_start|>' + message.role + '\\n' + content }}\n            {%- endif %}\n        {%- else %}\n            {{- '<|im_start|>' + message.role + '\\n' + content }}\n        {%- endif %}\n        {%- if message.tool_calls %}\n            {%- for tool_call in message.tool_calls %}\n                {%- if (loop.first and content) or (not loop.first) %}\n                    {{- '\\n' }}\n                {%- endif %}\n                {%- if tool_call.function %}\n                    {%- set tool_call = tool_call.function %}\n                {%- endif %}\n                {{- '<tool_call>\\n{\"name\": \"' }}\n                {{- tool_call.name }}\n                {{- '\", \"arguments\": ' }}\n                {%- if tool_call.arguments is string %}\n                    {{- tool_call.arguments }}\n                {%- else %}\n                    {{- tool_call.arguments | tojson }}\n                {%- endif %}\n                {{- '}\\n</tool_call>' }}\n            {%- endfor %}\n        {%- endif %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if loop.first or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n    {%- if enable_thinking is defined and enable_thinking is false %}\n        {{- '<think>\\n\\n</think>\\n\\n' }}\n    {%- endif %}\n{%- endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}