Add files using upload-large-folder tool

.gitattributes

@@ -33,3 +33,9 @@ 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
+checkpoint-1876/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-3752/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-938/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-2814/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-4690/tokenizer.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,61 @@
+---
+library_name: transformers
+license: other
+base_model: Qwen/Qwen3.5-9B-Base
+tags:
+- llama-factory
+- full
+- generated_from_trainer
+model-index:
+- name: omniscience_vqa_5epoch_full
+  results: []
+---
+
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+# omniscience_vqa_5epoch_full
+
+This model is a fine-tuned version of [/jizhicfs/wongzhenhao/models/Qwen3.5-9B-base](https://huggingface.co//jizhicfs/wongzhenhao/models/Qwen3.5-9B-base) on the omniscience_vqa dataset.
+
+## Model description
+
+More information needed
+
+## Intended uses & limitations
+
+More information needed
+
+## Training and evaluation data
+
+More information needed
+
+## Training procedure
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+- learning_rate: 1e-05
+- train_batch_size: 1
+- eval_batch_size: 8
+- seed: 42
+- distributed_type: multi-GPU
+- num_devices: 8
+- gradient_accumulation_steps: 8
+- total_train_batch_size: 64
+- total_eval_batch_size: 64
+- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
+- lr_scheduler_type: cosine
+- lr_scheduler_warmup_steps: 0.03
+- num_epochs: 5
+
+### Training results
+
+
+
+### Framework versions
+
+- Transformers 5.6.0
+- Pytorch 2.8.0+cu128
+- Datasets 4.0.0
+- Tokenizers 0.22.2

all_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 5.0,
+    "total_flos": 1.2342291469267108e+18,
+    "train_loss": 0.17693393766117502,
+    "train_runtime": 154131.4034,
+    "train_samples_per_second": 1.946,
+    "train_steps_per_second": 0.03
+}

chat_template.jinja

@@ -0,0 +1,154 @@
+{%- set image_count = namespace(value=0) %}
+{%- set video_count = namespace(value=0) %}
+{%- macro render_content(content, do_vision_count, is_system_content=false) %}
+    {%- if content is string %}
+        {{- content }}
+    {%- elif content is iterable and content is not mapping %}
+        {%- for item in content %}
+            {%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
+                {%- if is_system_content %}
+                    {{- raise_exception('System message cannot contain images.') }}
+                {%- endif %}
+                {%- if do_vision_count %}
+                    {%- set image_count.value = image_count.value + 1 %}
+                {%- endif %}
+                {%- if add_vision_id %}
+                    {{- 'Picture ' ~ image_count.value ~ ': ' }}
+                {%- endif %}
+                {{- '<|vision_start|><|image_pad|><|vision_end|>' }}
+            {%- elif 'video' in item or item.type == 'video' %}
+                {%- if is_system_content %}
+                    {{- raise_exception('System message cannot contain videos.') }}
+                {%- endif %}
+                {%- if do_vision_count %}
+                    {%- set video_count.value = video_count.value + 1 %}
+                {%- endif %}
+                {%- if add_vision_id %}
+                    {{- 'Video ' ~ video_count.value ~ ': ' }}
+                {%- endif %}
+                {{- '<|vision_start|><|video_pad|><|vision_end|>' }}
+            {%- elif 'text' in item %}
+                {{- item.text }}
+            {%- else %}
+                {{- raise_exception('Unexpected item type in content.') }}
+            {%- endif %}
+        {%- endfor %}
+    {%- elif content is none or content is undefined %}
+        {{- '' }}
+    {%- else %}
+        {{- raise_exception('Unexpected content type.') }}
+    {%- endif %}
+{%- endmacro %}
+{%- if not messages %}
+    {{- raise_exception('No messages provided.') }}
+{%- endif %}
+{%- if tools and tools is iterable and tools is not mapping %}
+    {{- '<|im_start|>system\n' }}
+    {{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>" }}
+    {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <tool_call></tool_call> XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n</IMPORTANT>' }}
+    {%- if messages[0].role == 'system' %}
+        {%- set content = render_content(messages[0].content, false, true)|trim %}
+        {%- if content %}
+            {{- '\n\n' + content }}
+        {%- endif %}
+    {%- endif %}
+    {{- '<|im_end|>\n' }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {%- set content = render_content(messages[0].content, false, true)|trim %}
+        {{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+    {%- set index = (messages|length - 1) - loop.index0 %}
+    {%- if ns.multi_step_tool and message.role == "user" %}
+        {%- set content = render_content(message.content, false)|trim %}
+        {%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
+            {%- set ns.multi_step_tool = false %}
+            {%- set ns.last_query_index = index %}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if ns.multi_step_tool %}
+    {{- raise_exception('No user query found in messages.') }}
+{%- endif %}
+{%- for message in messages %}
+    {%- set content = render_content(message.content, true)|trim %}
+    {%- if message.role == "system" %}
+        {%- if not loop.first %}
+            {{- raise_exception('System message must be at the beginning.') }}
+        {%- endif %}
+    {%- elif message.role == "user" %}
+        {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is string %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in content %}
+                {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+                {%- set content = content.split('</think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- set reasoning_content = reasoning_content|trim %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\n\n' + content }}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if tool_call.function is defined %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {%- if loop.first %}
+                    {%- if content|trim %}
+                        {{- '\n\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
+                    {%- else %}
+                        {{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
+                    {%- endif %}
+                {%- else %}
+                    {{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
+                {%- endif %}
+                {%- if tool_call.arguments is defined %}
+                    {%- for args_name, args_value in tool_call.arguments|items %}
+                        {{- '<parameter=' + args_name + '>\n' }}
+                        {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %}
+                        {{- args_value }}
+                        {{- '\n</parameter>\n' }}
+                    {%- endfor %}
+                {%- endif %}
+                {{- '</function>\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.previtem and loop.previtem.role != "tool" %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- content }}
+        {{- '\n</tool_response>' }}
+        {%- if not loop.last and loop.nextitem.role != "tool" %}
+            {{- '<|im_end|>\n' }}
+        {%- elif loop.last %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- else %}
+        {{- raise_exception('Unexpected message role.') }}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+    {%- if enable_thinking is defined and enable_thinking is false %}
+        {{- '<think>\n\n</think>\n\n' }}
+    {%- else %}
+        {{- '<think>\n' }}
+    {%- endif %}
+{%- endif %}

checkpoint-1876/chat_template.jinja

@@ -0,0 +1,154 @@
+{%- set image_count = namespace(value=0) %}
+{%- set video_count = namespace(value=0) %}
+{%- macro render_content(content, do_vision_count, is_system_content=false) %}
+    {%- if content is string %}
+        {{- content }}
+    {%- elif content is iterable and content is not mapping %}
+        {%- for item in content %}
+            {%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
+                {%- if is_system_content %}
+                    {{- raise_exception('System message cannot contain images.') }}
+                {%- endif %}
+                {%- if do_vision_count %}
+                    {%- set image_count.value = image_count.value + 1 %}
+                {%- endif %}
+                {%- if add_vision_id %}
+                    {{- 'Picture ' ~ image_count.value ~ ': ' }}
+                {%- endif %}
+                {{- '<|vision_start|><|image_pad|><|vision_end|>' }}
+            {%- elif 'video' in item or item.type == 'video' %}
+                {%- if is_system_content %}
+                    {{- raise_exception('System message cannot contain videos.') }}
+                {%- endif %}
+                {%- if do_vision_count %}
+                    {%- set video_count.value = video_count.value + 1 %}
+                {%- endif %}
+                {%- if add_vision_id %}
+                    {{- 'Video ' ~ video_count.value ~ ': ' }}
+                {%- endif %}
+                {{- '<|vision_start|><|video_pad|><|vision_end|>' }}
+            {%- elif 'text' in item %}
+                {{- item.text }}
+            {%- else %}
+                {{- raise_exception('Unexpected item type in content.') }}
+            {%- endif %}
+        {%- endfor %}
+    {%- elif content is none or content is undefined %}
+        {{- '' }}
+    {%- else %}
+        {{- raise_exception('Unexpected content type.') }}
+    {%- endif %}
+{%- endmacro %}
+{%- if not messages %}
+    {{- raise_exception('No messages provided.') }}
+{%- endif %}
+{%- if tools and tools is iterable and tools is not mapping %}
+    {{- '<|im_start|>system\n' }}
+    {{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>" }}
+    {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <tool_call></tool_call> XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n</IMPORTANT>' }}
+    {%- if messages[0].role == 'system' %}
+        {%- set content = render_content(messages[0].content, false, true)|trim %}
+        {%- if content %}
+            {{- '\n\n' + content }}
+        {%- endif %}
+    {%- endif %}
+    {{- '<|im_end|>\n' }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {%- set content = render_content(messages[0].content, false, true)|trim %}
+        {{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+    {%- set index = (messages|length - 1) - loop.index0 %}
+    {%- if ns.multi_step_tool and message.role == "user" %}
+        {%- set content = render_content(message.content, false)|trim %}
+        {%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
+            {%- set ns.multi_step_tool = false %}
+            {%- set ns.last_query_index = index %}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if ns.multi_step_tool %}
+    {{- raise_exception('No user query found in messages.') }}
+{%- endif %}
+{%- for message in messages %}
+    {%- set content = render_content(message.content, true)|trim %}
+    {%- if message.role == "system" %}
+        {%- if not loop.first %}
+            {{- raise_exception('System message must be at the beginning.') }}
+        {%- endif %}
+    {%- elif message.role == "user" %}
+        {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is string %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in content %}
+                {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+                {%- set content = content.split('</think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- set reasoning_content = reasoning_content|trim %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\n\n' + content }}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if tool_call.function is defined %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {%- if loop.first %}
+                    {%- if content|trim %}
+                        {{- '\n\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
+                    {%- else %}
+                        {{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
+                    {%- endif %}
+                {%- else %}
+                    {{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
+                {%- endif %}
+                {%- if tool_call.arguments is defined %}
+                    {%- for args_name, args_value in tool_call.arguments|items %}
+                        {{- '<parameter=' + args_name + '>\n' }}
+                        {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %}
+                        {{- args_value }}
+                        {{- '\n</parameter>\n' }}
+                    {%- endfor %}
+                {%- endif %}
+                {{- '</function>\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.previtem and loop.previtem.role != "tool" %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- content }}
+        {{- '\n</tool_response>' }}
+        {%- if not loop.last and loop.nextitem.role != "tool" %}
+            {{- '<|im_end|>\n' }}
+        {%- elif loop.last %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- else %}
+        {{- raise_exception('Unexpected message role.') }}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+    {%- if enable_thinking is defined and enable_thinking is false %}
+        {{- '<think>\n\n</think>\n\n' }}
+    {%- else %}
+        {{- '<think>\n' }}
+    {%- endif %}
+{%- endif %}

checkpoint-1876/config.json

@@ -0,0 +1,113 @@
+{
+  "architectures": [
+    "Qwen3_5ForConditionalGeneration"
+  ],
+  "dtype": "bfloat16",
+  "eos_token_id": 248046,
+  "hidden_size": 4096,
+  "image_token_id": 248056,
+  "model_type": "qwen3_5",
+  "pad_token_id": 248044,
+  "text_config": {
+    "attention_bias": false,
+    "attention_dropout": 0.0,
+    "attn_output_gate": true,
+    "bos_token_id": null,
+    "dtype": "bfloat16",
+    "eos_token_id": 248044,
+    "full_attention_interval": 4,
+    "head_dim": 256,
+    "hidden_act": "silu",
+    "hidden_size": 4096,
+    "initializer_range": 0.02,
+    "intermediate_size": 12288,
+    "layer_types": [
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention",
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention",
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention",
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention",
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention",
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention",
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention",
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention"
+    ],
+    "linear_conv_kernel_dim": 4,
+    "linear_key_head_dim": 128,
+    "linear_num_key_heads": 16,
+    "linear_num_value_heads": 32,
+    "linear_value_head_dim": 128,
+    "mamba_ssm_dtype": "float32",
+    "max_position_embeddings": 262144,
+    "mlp_only_layers": [],
+    "model_type": "qwen3_5_text",
+    "mtp_num_hidden_layers": 1,
+    "mtp_use_dedicated_embeddings": false,
+    "num_attention_heads": 16,
+    "num_hidden_layers": 32,
+    "num_key_value_heads": 4,
+    "pad_token_id": null,
+    "partial_rotary_factor": 0.25,
+    "rms_norm_eps": 1e-06,
+    "rope_parameters": {
+      "mrope_interleaved": true,
+      "mrope_section": [
+        11,
+        11,
+        10
+      ],
+      "partial_rotary_factor": 0.25,
+      "rope_theta": 10000000,
+      "rope_type": "default"
+    },
+    "tie_word_embeddings": false,
+    "use_cache": false,
+    "vocab_size": 248320
+  },
+  "tie_word_embeddings": false,
+  "transformers_version": "5.6.0",
+  "use_cache": false,
+  "video_token_id": 248057,
+  "vision_config": {
+    "deepstack_visual_indexes": [],
+    "depth": 27,
+    "dtype": "bfloat16",
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 1152,
+    "in_channels": 3,
+    "initializer_range": 0.02,
+    "intermediate_size": 4304,
+    "model_type": "qwen3_5_vision",
+    "num_heads": 16,
+    "num_position_embeddings": 2304,
+    "out_hidden_size": 4096,
+    "patch_size": 16,
+    "spatial_merge_size": 2,
+    "temporal_patch_size": 2
+  },
+  "vision_end_token_id": 248054,
+  "vision_start_token_id": 248053
+}

checkpoint-1876/generation_config.json

@@ -0,0 +1,10 @@
+{
+  "_from_model_config": true,
+  "eos_token_id": [
+    248046,
+    248044
+  ],
+  "pad_token_id": 248044,
+  "transformers_version": "5.6.0",
+  "use_cache": true
+}

checkpoint-1876/global_step1876/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bf82b1a19ccba35f81e8e4b185af29d96062602142701d1f6a58655ec37f66b0
+size 13430710881

checkpoint-1876/global_step1876/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:859b6d2114bedd7b24d6f358394c46432690ce5062ebc37252e35c3971d45336
+size 13430710881

checkpoint-1876/global_step1876/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:282e102cd8975f730d3586fda0567cc6e8f5fe082af8e6965cb913f0aeac067f
+size 13430710881

checkpoint-1876/global_step1876/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3a58e790c156a3693fba2dac4ef74d8c6aa2c4647eab22086e62a7ba10395e80
+size 13430710881

checkpoint-1876/global_step1876/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:674c6e8db3ee3b2b9900936f1994fcf2a7fce53015021061b47b31349f4d94ae
+size 13430710881

checkpoint-1876/global_step1876/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e00b29cafafcdd5d63ee1b31e9b1634a44ebf6abdd5d749ea4bc1cfeeda0d280
+size 13430710881

checkpoint-1876/global_step1876/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:76bac6ebff4f2a5d30b7b09c386db322c7e79642ad05c3f54188dc3917dfd1da
+size 13430710881

checkpoint-1876/global_step1876/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f04f9806a0f91f048655c461d8ccc57e97c0a625faa734e1bcfe8a07f1c79876
+size 13430710881

checkpoint-1876/global_step1876/zero_pp_rank_0_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0d63a80ace95e5005f5b156c78e316b02d3989f29a108b7a1b95f9194ea53cde
+size 114502951

checkpoint-1876/global_step1876/zero_pp_rank_1_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7bcd78952c577e661cc8fd8f4a6468f0d0c08bffaaebb1cf7ed8b93c9a662300
+size 114502951

checkpoint-1876/global_step1876/zero_pp_rank_2_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:893d997e54e25165ae787b6e4a2ffcf4cf150c478f77aca79ac0a1766d4a5fa2
+size 114502951

checkpoint-1876/global_step1876/zero_pp_rank_3_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fba79de4f7e65c1d0163a10753b916724ff2e30b646ec2d2ec4c96cd58f4ef82
+size 114502951

checkpoint-1876/global_step1876/zero_pp_rank_4_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3be3f825fdd4eb2c1f74c08705f157cdf6b843889882e8d01d87ddc148ecbae8
+size 114502951

checkpoint-1876/global_step1876/zero_pp_rank_5_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a0d91b67f4f5b909e1050aa7ec5ce71d4b9cf8004b53a9c7c62f649d073b6c6d
+size 114502951

checkpoint-1876/global_step1876/zero_pp_rank_6_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f064dac5a20fdfd0ddeff2edc31fbaef5305e3a192f0630a00f241513143f205
+size 114502951

checkpoint-1876/global_step1876/zero_pp_rank_7_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8634541095796bb8b4accb4395a454df7df547bab8136b6801a1efdf9f908996
+size 114502951

checkpoint-1876/latest

@@ -0,0 +1 @@
+global_step1876

checkpoint-1876/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:38b9ffdc6f38e9d6411ec7e8e8463a9768b88b15207291b5fb8655b7f27adeb6
+size 18819722392

checkpoint-1876/processor_config.json

@@ -0,0 +1,60 @@
+{
+  "image_processor": {
+    "do_convert_rgb": true,
+    "do_normalize": true,
+    "do_rescale": true,
+    "do_resize": true,
+    "image_mean": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "image_processor_type": "Qwen2VLImageProcessor",
+    "image_std": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "merge_size": 2,
+    "patch_size": 16,
+    "resample": 3,
+    "rescale_factor": 0.00392156862745098,
+    "size": {
+      "longest_edge": 16777216,
+      "shortest_edge": 65536
+    },
+    "temporal_patch_size": 2
+  },
+  "processor_class": "Qwen3VLProcessor",
+  "video_processor": {
+    "do_convert_rgb": true,
+    "do_normalize": true,
+    "do_rescale": true,
+    "do_resize": true,
+    "do_sample_frames": true,
+    "fps": 2,
+    "image_mean": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "image_std": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "max_frames": 768,
+    "merge_size": 2,
+    "min_frames": 4,
+    "patch_size": 16,
+    "resample": 3,
+    "rescale_factor": 0.00392156862745098,
+    "return_metadata": false,
+    "size": {
+      "longest_edge": 234881024,
+      "shortest_edge": 4096
+    },
+    "temporal_patch_size": 2,
+    "video_processor_type": "Qwen3VLVideoProcessor"
+  }
+}

checkpoint-1876/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8bfe1981024ef92f2da08a90c72c7c793d1cc9de1547abd2556c968be70232eb
+size 16389

checkpoint-1876/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a35b845d476d830805793c3dcf8ac2daad87fec289bff3f7eda9e72fc374eda1
+size 16389

checkpoint-1876/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:03e9880996b01262a807d1ec3ebd91eee540e08130a14a45a4648731fd0d48a9
+size 16389

checkpoint-1876/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ee25c237d6fe62ec76adcf7daf899d7ed32eab5d1a5b447b911f4451c9a1b258
+size 16389

checkpoint-1876/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7a6b31133f29a8fc0cb538aa807d6a403bd51939336bfd425cd3d122d8c5595c
+size 16389

checkpoint-1876/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a26c55b5c7fa0522b1d27b2c00a7ea77ad010f19a1321991165c5c972b8fa97a
+size 16389

checkpoint-1876/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4a1a3cf85626196804f25a8293e22dc561bba068a70fb123e04afe4896c33972
+size 16389

checkpoint-1876/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:28f87c1ee5f5db346c7b913137cbccd196eaf8ec5a4cf9f192418a3069269b49
+size 16389

checkpoint-1876/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dced6746ac599c81802139f61ed3e1504bc60c84f0fbf29558268f32b322bbbf
+size 1465

checkpoint-1876/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:06b9509352d2af50381ab2247e083b80d32d5c0aba91c272ca9ff729b6a0e523
+size 19989325

checkpoint-1876/tokenizer_config.json

@@ -0,0 +1,34 @@
+{
+  "add_prefix_space": false,
+  "audio_bos_token": "<|audio_start|>",
+  "audio_eos_token": "<|audio_end|>",
+  "audio_token": "<|audio_pad|>",
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "image_token": "<|image_pad|>",
+  "is_local": true,
+  "local_files_only": false,
+  "model_max_length": 262144,
+  "model_specific_special_tokens": {
+    "audio_bos_token": "<|audio_start|>",
+    "audio_eos_token": "<|audio_end|>",
+    "audio_token": "<|audio_pad|>",
+    "image_token": "<|image_pad|>",
+    "video_token": "<|video_pad|>",
+    "vision_bos_token": "<|vision_start|>",
+    "vision_eos_token": "<|vision_end|>"
+  },
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
+  "processor_class": "Qwen3VLProcessor",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null,
+  "video_token": "<|video_pad|>",
+  "vision_bos_token": "<|vision_start|>",
+  "vision_eos_token": "<|vision_end|>"
+}

checkpoint-1876/trainer_state.json

@@ -0,0 +1,1343 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.0,
+  "eval_steps": 500,
+  "global_step": 1876,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.010666666666666666,
+      "grad_norm": 14.719677069033247,
+      "learning_rate": 6.382978723404255e-07,
+      "loss": 1.131131935119629,
+      "step": 10
+    },
+    {
+      "epoch": 0.021333333333333333,
+      "grad_norm": 19.789131904372997,
+      "learning_rate": 1.347517730496454e-06,
+      "loss": 0.8277447700500489,
+      "step": 20
+    },
+    {
+      "epoch": 0.032,
+      "grad_norm": 4.008811801062304,
+      "learning_rate": 2.0567375886524823e-06,
+      "loss": 0.5474711418151855,
+      "step": 30
+    },
+    {
+      "epoch": 0.042666666666666665,
+      "grad_norm": 3.4969831449617863,
+      "learning_rate": 2.765957446808511e-06,
+      "loss": 0.4969797611236572,
+      "step": 40
+    },
+    {
+      "epoch": 0.05333333333333334,
+      "grad_norm": 3.496710211411394,
+      "learning_rate": 3.4751773049645393e-06,
+      "loss": 0.4998605728149414,
+      "step": 50
+    },
+    {
+      "epoch": 0.064,
+      "grad_norm": 3.3031965730768413,
+      "learning_rate": 4.184397163120568e-06,
+      "loss": 0.4771441459655762,
+      "step": 60
+    },
+    {
+      "epoch": 0.07466666666666667,
+      "grad_norm": 3.5681289223245787,
+      "learning_rate": 4.893617021276596e-06,
+      "loss": 0.4836556434631348,
+      "step": 70
+    },
+    {
+      "epoch": 0.08533333333333333,
+      "grad_norm": 3.2287114465847284,
+      "learning_rate": 5.602836879432625e-06,
+      "loss": 0.47313661575317384,
+      "step": 80
+    },
+    {
+      "epoch": 0.096,
+      "grad_norm": 2.9987978137542015,
+      "learning_rate": 6.312056737588653e-06,
+      "loss": 0.4791095733642578,
+      "step": 90
+    },
+    {
+      "epoch": 0.10666666666666667,
+      "grad_norm": 3.340378081154499,
+      "learning_rate": 7.021276595744682e-06,
+      "loss": 0.47806363105773925,
+      "step": 100
+    },
+    {
+      "epoch": 0.11733333333333333,
+      "grad_norm": 2.7047692208810417,
+      "learning_rate": 7.73049645390071e-06,
+      "loss": 0.49144439697265624,
+      "step": 110
+    },
+    {
+      "epoch": 0.128,
+      "grad_norm": 2.9034932565546754,
+      "learning_rate": 8.439716312056738e-06,
+      "loss": 0.4856122970581055,
+      "step": 120
+    },
+    {
+      "epoch": 0.13866666666666666,
+      "grad_norm": 2.8492532284524983,
+      "learning_rate": 9.148936170212767e-06,
+      "loss": 0.49585614204406736,
+      "step": 130
+    },
+    {
+      "epoch": 0.14933333333333335,
+      "grad_norm": 3.1897219221657913,
+      "learning_rate": 9.858156028368795e-06,
+      "loss": 0.4963822364807129,
+      "step": 140
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 3.180291054321027,
+      "learning_rate": 9.999923689069412e-06,
+      "loss": 0.49301981925964355,
+      "step": 150
+    },
+    {
+      "epoch": 0.17066666666666666,
+      "grad_norm": 2.5896623193472585,
+      "learning_rate": 9.999613679906083e-06,
+      "loss": 0.4908170223236084,
+      "step": 160
+    },
+    {
+      "epoch": 0.18133333333333335,
+      "grad_norm": 2.582517102244016,
+      "learning_rate": 9.999065217851136e-06,
+      "loss": 0.512379503250122,
+      "step": 170
+    },
+    {
+      "epoch": 0.192,
+      "grad_norm": 2.9481033540091173,
+      "learning_rate": 9.998278329063063e-06,
+      "loss": 0.5089564323425293,
+      "step": 180
+    },
+    {
+      "epoch": 0.20266666666666666,
+      "grad_norm": 2.742530929422462,
+      "learning_rate": 9.997253051071943e-06,
+      "loss": 0.49326133728027344,
+      "step": 190
+    },
+    {
+      "epoch": 0.21333333333333335,
+      "grad_norm": 2.765768124507322,
+      "learning_rate": 9.99598943277766e-06,
+      "loss": 0.4973162651062012,
+      "step": 200
+    },
+    {
+      "epoch": 0.224,
+      "grad_norm": 2.770226206469208,
+      "learning_rate": 9.994487534447554e-06,
+      "loss": 0.49165959358215333,
+      "step": 210
+    },
+    {
+      "epoch": 0.23466666666666666,
+      "grad_norm": 2.6739784925959724,
+      "learning_rate": 9.992747427713561e-06,
+      "loss": 0.5012942314147949,
+      "step": 220
+    },
+    {
+      "epoch": 0.24533333333333332,
+      "grad_norm": 2.564165673280976,
+      "learning_rate": 9.990769195568787e-06,
+      "loss": 0.4979718208312988,
+      "step": 230
+    },
+    {
+      "epoch": 0.256,
+      "grad_norm": 2.5032814386910713,
+      "learning_rate": 9.988552932363563e-06,
+      "loss": 0.5078562736511231,
+      "step": 240
+    },
+    {
+      "epoch": 0.26666666666666666,
+      "grad_norm": 3.0170101170042805,
+      "learning_rate": 9.986098743800922e-06,
+      "loss": 0.5044580936431885,
+      "step": 250
+    },
+    {
+      "epoch": 0.2773333333333333,
+      "grad_norm": 2.679278781269865,
+      "learning_rate": 9.983406746931588e-06,
+      "loss": 0.515423583984375,
+      "step": 260
+    },
+    {
+      "epoch": 0.288,
+      "grad_norm": 2.446831637429734,
+      "learning_rate": 9.980477070148363e-06,
+      "loss": 0.4906794548034668,
+      "step": 270
+    },
+    {
+      "epoch": 0.2986666666666667,
+      "grad_norm": 2.4600457457453326,
+      "learning_rate": 9.977309853180024e-06,
+      "loss": 0.4913760185241699,
+      "step": 280
+    },
+    {
+      "epoch": 0.30933333333333335,
+      "grad_norm": 2.6584767292187426,
+      "learning_rate": 9.973905247084654e-06,
+      "loss": 0.4985548496246338,
+      "step": 290
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 2.4646094466368766,
+      "learning_rate": 9.970263414242429e-06,
+      "loss": 0.49314355850219727,
+      "step": 300
+    },
+    {
+      "epoch": 0.33066666666666666,
+      "grad_norm": 3.149315472839174,
+      "learning_rate": 9.966384528347886e-06,
+      "loss": 0.5047734260559082,
+      "step": 310
+    },
+    {
+      "epoch": 0.3413333333333333,
+      "grad_norm": 2.618183661885374,
+      "learning_rate": 9.96226877440163e-06,
+      "loss": 0.4957937240600586,
+      "step": 320
+    },
+    {
+      "epoch": 0.352,
+      "grad_norm": 2.615669163983153,
+      "learning_rate": 9.957916348701515e-06,
+      "loss": 0.49486532211303713,
+      "step": 330
+    },
+    {
+      "epoch": 0.3626666666666667,
+      "grad_norm": 2.6327990047828793,
+      "learning_rate": 9.953327458833276e-06,
+      "loss": 0.49072465896606443,
+      "step": 340
+    },
+    {
+      "epoch": 0.37333333333333335,
+      "grad_norm": 2.8831237701931713,
+      "learning_rate": 9.948502323660635e-06,
+      "loss": 0.5094201564788818,
+      "step": 350
+    },
+    {
+      "epoch": 0.384,
+      "grad_norm": 2.4981375259751206,
+      "learning_rate": 9.943441173314865e-06,
+      "loss": 0.4819591522216797,
+      "step": 360
+    },
+    {
+      "epoch": 0.39466666666666667,
+      "grad_norm": 2.3457990842275374,
+      "learning_rate": 9.938144249183797e-06,
+      "loss": 0.507014274597168,
+      "step": 370
+    },
+    {
+      "epoch": 0.4053333333333333,
+      "grad_norm": 2.4847312856992176,
+      "learning_rate": 9.932611803900328e-06,
+      "loss": 0.48714618682861327,
+      "step": 380
+    },
+    {
+      "epoch": 0.416,
+      "grad_norm": 2.3913464084324945,
+      "learning_rate": 9.926844101330362e-06,
+      "loss": 0.4863110542297363,
+      "step": 390
+    },
+    {
+      "epoch": 0.4266666666666667,
+      "grad_norm": 2.652794886653921,
+      "learning_rate": 9.920841416560221e-06,
+      "loss": 0.49074840545654297,
+      "step": 400
+    },
+    {
+      "epoch": 0.43733333333333335,
+      "grad_norm": 2.5625986892418737,
+      "learning_rate": 9.914604035883537e-06,
+      "loss": 0.4913002014160156,
+      "step": 410
+    },
+    {
+      "epoch": 0.448,
+      "grad_norm": 2.4680795319586846,
+      "learning_rate": 9.908132256787586e-06,
+      "loss": 0.48008413314819337,
+      "step": 420
+    },
+    {
+      "epoch": 0.45866666666666667,
+      "grad_norm": 2.446518908575631,
+      "learning_rate": 9.901426387939104e-06,
+      "loss": 0.484285306930542,
+      "step": 430
+    },
+    {
+      "epoch": 0.4693333333333333,
+      "grad_norm": 2.3434078549751973,
+      "learning_rate": 9.894486749169564e-06,
+      "loss": 0.48981361389160155,
+      "step": 440
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 2.5243663001591665,
+      "learning_rate": 9.887313671459929e-06,
+      "loss": 0.4931126594543457,
+      "step": 450
+    },
+    {
+      "epoch": 0.49066666666666664,
+      "grad_norm": 2.5905415730646935,
+      "learning_rate": 9.879907496924853e-06,
+      "loss": 0.4852896690368652,
+      "step": 460
+    },
+    {
+      "epoch": 0.5013333333333333,
+      "grad_norm": 2.3175960946503635,
+      "learning_rate": 9.872268578796377e-06,
+      "loss": 0.4794930458068848,
+      "step": 470
+    },
+    {
+      "epoch": 0.512,
+      "grad_norm": 2.3169110867720084,
+      "learning_rate": 9.864397281407074e-06,
+      "loss": 0.49527502059936523,
+      "step": 480
+    },
+    {
+      "epoch": 0.5226666666666666,
+      "grad_norm": 2.5081805539045763,
+      "learning_rate": 9.856293980172676e-06,
+      "loss": 0.49238901138305663,
+      "step": 490
+    },
+    {
+      "epoch": 0.5333333333333333,
+      "grad_norm": 2.4148307543494756,
+      "learning_rate": 9.847959061574163e-06,
+      "loss": 0.48439202308654783,
+      "step": 500
+    },
+    {
+      "epoch": 0.544,
+      "grad_norm": 2.417268129552713,
+      "learning_rate": 9.83939292313934e-06,
+      "loss": 0.48358449935913084,
+      "step": 510
+    },
+    {
+      "epoch": 0.5546666666666666,
+      "grad_norm": 2.4212977880426583,
+      "learning_rate": 9.830595973423868e-06,
+      "loss": 0.4881169319152832,
+      "step": 520
+    },
+    {
+      "epoch": 0.5653333333333334,
+      "grad_norm": 2.2373251511625227,
+      "learning_rate": 9.821568631991785e-06,
+      "loss": 0.4937411308288574,
+      "step": 530
+    },
+    {
+      "epoch": 0.576,
+      "grad_norm": 2.6067848093793113,
+      "learning_rate": 9.812311329395492e-06,
+      "loss": 0.4851250648498535,
+      "step": 540
+    },
+    {
+      "epoch": 0.5866666666666667,
+      "grad_norm": 2.505779269023947,
+      "learning_rate": 9.802824507155215e-06,
+      "loss": 0.476285457611084,
+      "step": 550
+    },
+    {
+      "epoch": 0.5973333333333334,
+      "grad_norm": 2.453798365354759,
+      "learning_rate": 9.793108617737953e-06,
+      "loss": 0.4898886203765869,
+      "step": 560
+    },
+    {
+      "epoch": 0.608,
+      "grad_norm": 2.336419532943086,
+      "learning_rate": 9.783164124535896e-06,
+      "loss": 0.47934665679931643,
+      "step": 570
+    },
+    {
+      "epoch": 0.6186666666666667,
+      "grad_norm": 2.3622052222837984,
+      "learning_rate": 9.77299150184432e-06,
+      "loss": 0.4934229850769043,
+      "step": 580
+    },
+    {
+      "epoch": 0.6293333333333333,
+      "grad_norm": 2.355971888227553,
+      "learning_rate": 9.762591234838972e-06,
+      "loss": 0.4741045951843262,
+      "step": 590
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 2.335087591202486,
+      "learning_rate": 9.751963819552925e-06,
+      "loss": 0.4835971832275391,
+      "step": 600
+    },
+    {
+      "epoch": 0.6506666666666666,
+      "grad_norm": 2.312226806890674,
+      "learning_rate": 9.741109762852924e-06,
+      "loss": 0.4975107669830322,
+      "step": 610
+    },
+    {
+      "epoch": 0.6613333333333333,
+      "grad_norm": 2.4120574289328833,
+      "learning_rate": 9.730029582415209e-06,
+      "loss": 0.4812762260437012,
+      "step": 620
+    },
+    {
+      "epoch": 0.672,
+      "grad_norm": 2.70147229008976,
+      "learning_rate": 9.718723806700818e-06,
+      "loss": 0.46732497215270996,
+      "step": 630
+    },
+    {
+      "epoch": 0.6826666666666666,
+      "grad_norm": 2.4120813404749573,
+      "learning_rate": 9.707192974930403e-06,
+      "loss": 0.48064889907836916,
+      "step": 640
+    },
+    {
+      "epoch": 0.6933333333333334,
+      "grad_norm": 2.3262098681526093,
+      "learning_rate": 9.695437637058487e-06,
+      "loss": 0.49033608436584475,
+      "step": 650
+    },
+    {
+      "epoch": 0.704,
+      "grad_norm": 2.6101360877178346,
+      "learning_rate": 9.68345835374725e-06,
+      "loss": 0.48049345016479494,
+      "step": 660
+    },
+    {
+      "epoch": 0.7146666666666667,
+      "grad_norm": 2.38416338086165,
+      "learning_rate": 9.67125569633979e-06,
+      "loss": 0.48430418968200684,
+      "step": 670
+    },
+    {
+      "epoch": 0.7253333333333334,
+      "grad_norm": 2.396147340040666,
+      "learning_rate": 9.658830246832858e-06,
+      "loss": 0.486968994140625,
+      "step": 680
+    },
+    {
+      "epoch": 0.736,
+      "grad_norm": 2.1880846756162775,
+      "learning_rate": 9.646182597849119e-06,
+      "loss": 0.46998023986816406,
+      "step": 690
+    },
+    {
+      "epoch": 0.7466666666666667,
+      "grad_norm": 2.3991316493747403,
+      "learning_rate": 9.633313352608871e-06,
+      "loss": 0.4902526378631592,
+      "step": 700
+    },
+    {
+      "epoch": 0.7573333333333333,
+      "grad_norm": 2.546776822869976,
+      "learning_rate": 9.620223124901289e-06,
+      "loss": 0.4668087959289551,
+      "step": 710
+    },
+    {
+      "epoch": 0.768,
+      "grad_norm": 2.186701096003848,
+      "learning_rate": 9.606912539055141e-06,
+      "loss": 0.48386144638061523,
+      "step": 720
+    },
+    {
+      "epoch": 0.7786666666666666,
+      "grad_norm": 2.28344654961423,
+      "learning_rate": 9.593382229909012e-06,
+      "loss": 0.4883030891418457,
+      "step": 730
+    },
+    {
+      "epoch": 0.7893333333333333,
+      "grad_norm": 2.289876953573637,
+      "learning_rate": 9.579632842781033e-06,
+      "loss": 0.4781497001647949,
+      "step": 740
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 2.2868985326044813,
+      "learning_rate": 9.565665033438089e-06,
+      "loss": 0.4757716655731201,
+      "step": 750
+    },
+    {
+      "epoch": 0.8106666666666666,
+      "grad_norm": 2.30563701368124,
+      "learning_rate": 9.551479468064565e-06,
+      "loss": 0.48414173126220705,
+      "step": 760
+    },
+    {
+      "epoch": 0.8213333333333334,
+      "grad_norm": 2.3272021596931296,
+      "learning_rate": 9.537076823230547e-06,
+      "loss": 0.4821342468261719,
+      "step": 770
+    },
+    {
+      "epoch": 0.832,
+      "grad_norm": 2.356076098592652,
+      "learning_rate": 9.522457785859572e-06,
+      "loss": 0.48393640518188474,
+      "step": 780
+    },
+    {
+      "epoch": 0.8426666666666667,
+      "grad_norm": 2.283081447884378,
+      "learning_rate": 9.507623053195861e-06,
+      "loss": 0.4777627944946289,
+      "step": 790
+    },
+    {
+      "epoch": 0.8533333333333334,
+      "grad_norm": 1.9593766326717093,
+      "learning_rate": 9.492573332771056e-06,
+      "loss": 0.4635260581970215,
+      "step": 800
+    },
+    {
+      "epoch": 0.864,
+      "grad_norm": 2.4282608759061093,
+      "learning_rate": 9.477309342370491e-06,
+      "loss": 0.48109025955200196,
+      "step": 810
+    },
+    {
+      "epoch": 0.8746666666666667,
+      "grad_norm": 2.2037180310212676,
+      "learning_rate": 9.461831809998944e-06,
+      "loss": 0.47229413986206054,
+      "step": 820
+    },
+    {
+      "epoch": 0.8853333333333333,
+      "grad_norm": 2.3422653774041384,
+      "learning_rate": 9.446141473845917e-06,
+      "loss": 0.47471017837524415,
+      "step": 830
+    },
+    {
+      "epoch": 0.896,
+      "grad_norm": 2.307529643267961,
+      "learning_rate": 9.430239082250438e-06,
+      "loss": 0.47556605339050295,
+      "step": 840
+    },
+    {
+      "epoch": 0.9066666666666666,
+      "grad_norm": 2.4685203961814,
+      "learning_rate": 9.41412539366536e-06,
+      "loss": 0.460905122756958,
+      "step": 850
+    },
+    {
+      "epoch": 0.9173333333333333,
+      "grad_norm": 2.253282348951304,
+      "learning_rate": 9.39780117662119e-06,
+      "loss": 0.46721577644348145,
+      "step": 860
+    },
+    {
+      "epoch": 0.928,
+      "grad_norm": 2.6725399753463663,
+      "learning_rate": 9.381267209689433e-06,
+      "loss": 0.4595684051513672,
+      "step": 870
+    },
+    {
+      "epoch": 0.9386666666666666,
+      "grad_norm": 2.1190030812596174,
+      "learning_rate": 9.364524281445464e-06,
+      "loss": 0.4708869934082031,
+      "step": 880
+    },
+    {
+      "epoch": 0.9493333333333334,
+      "grad_norm": 2.6687501945392844,
+      "learning_rate": 9.347573190430912e-06,
+      "loss": 0.4814455032348633,
+      "step": 890
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 2.293302659426518,
+      "learning_rate": 9.330414745115576e-06,
+      "loss": 0.46994471549987793,
+      "step": 900
+    },
+    {
+      "epoch": 0.9706666666666667,
+      "grad_norm": 2.3844259241813273,
+      "learning_rate": 9.31304976385887e-06,
+      "loss": 0.47742176055908203,
+      "step": 910
+    },
+    {
+      "epoch": 0.9813333333333333,
+      "grad_norm": 2.2795793162673776,
+      "learning_rate": 9.295479074870777e-06,
+      "loss": 0.4683531761169434,
+      "step": 920
+    },
+    {
+      "epoch": 0.992,
+      "grad_norm": 2.3912968167883277,
+      "learning_rate": 9.27770351617237e-06,
+      "loss": 0.4541953086853027,
+      "step": 930
+    },
+    {
+      "epoch": 1.0021333333333333,
+      "grad_norm": 1.7527643233253156,
+      "learning_rate": 9.259723935555824e-06,
+      "loss": 0.422662353515625,
+      "step": 940
+    },
+    {
+      "epoch": 1.0128,
+      "grad_norm": 2.3671794408283113,
+      "learning_rate": 9.241541190543995e-06,
+      "loss": 0.2570992946624756,
+      "step": 950
+    },
+    {
+      "epoch": 1.0234666666666667,
+      "grad_norm": 2.1688535112637095,
+      "learning_rate": 9.223156148349511e-06,
+      "loss": 0.25190885066986085,
+      "step": 960
+    },
+    {
+      "epoch": 1.0341333333333333,
+      "grad_norm": 2.2112826371708976,
+      "learning_rate": 9.204569685833416e-06,
+      "loss": 0.2573091983795166,
+      "step": 970
+    },
+    {
+      "epoch": 1.0448,
+      "grad_norm": 2.360134745847926,
+      "learning_rate": 9.185782689463352e-06,
+      "loss": 0.24791259765625,
+      "step": 980
+    },
+    {
+      "epoch": 1.0554666666666668,
+      "grad_norm": 2.1230653101557726,
+      "learning_rate": 9.166796055271267e-06,
+      "loss": 0.24686360359191895,
+      "step": 990
+    },
+    {
+      "epoch": 1.0661333333333334,
+      "grad_norm": 2.179214976563206,
+      "learning_rate": 9.147610688810695e-06,
+      "loss": 0.24492282867431642,
+      "step": 1000
+    },
+    {
+      "epoch": 1.0768,
+      "grad_norm": 2.0319612229736013,
+      "learning_rate": 9.128227505113557e-06,
+      "loss": 0.23637039661407472,
+      "step": 1010
+    },
+    {
+      "epoch": 1.0874666666666666,
+      "grad_norm": 2.145879982585577,
+      "learning_rate": 9.108647428646522e-06,
+      "loss": 0.244051194190979,
+      "step": 1020
+    },
+    {
+      "epoch": 1.0981333333333334,
+      "grad_norm": 2.114387055100977,
+      "learning_rate": 9.088871393266912e-06,
+      "loss": 0.2529242277145386,
+      "step": 1030
+    },
+    {
+      "epoch": 1.1088,
+      "grad_norm": 2.035177482973937,
+      "learning_rate": 9.068900342178169e-06,
+      "loss": 0.2405726671218872,
+      "step": 1040
+    },
+    {
+      "epoch": 1.1194666666666666,
+      "grad_norm": 2.2418308787896217,
+      "learning_rate": 9.048735227884861e-06,
+      "loss": 0.24377832412719727,
+      "step": 1050
+    },
+    {
+      "epoch": 1.1301333333333332,
+      "grad_norm": 2.34726411543382,
+      "learning_rate": 9.028377012147262e-06,
+      "loss": 0.24974491596221923,
+      "step": 1060
+    },
+    {
+      "epoch": 1.1408,
+      "grad_norm": 2.3290056204741427,
+      "learning_rate": 9.007826665935477e-06,
+      "loss": 0.23754820823669434,
+      "step": 1070
+    },
+    {
+      "epoch": 1.1514666666666666,
+      "grad_norm": 2.0722263152081863,
+      "learning_rate": 8.987085169383125e-06,
+      "loss": 0.24723689556121825,
+      "step": 1080
+    },
+    {
+      "epoch": 1.1621333333333332,
+      "grad_norm": 2.201847254257956,
+      "learning_rate": 8.966153511740612e-06,
+      "loss": 0.2497556686401367,
+      "step": 1090
+    },
+    {
+      "epoch": 1.1728,
+      "grad_norm": 2.189179516457379,
+      "learning_rate": 8.945032691327925e-06,
+      "loss": 0.24566569328308105,
+      "step": 1100
+    },
+    {
+      "epoch": 1.1834666666666667,
+      "grad_norm": 2.3721350824712304,
+      "learning_rate": 8.923723715487038e-06,
+      "loss": 0.25868451595306396,
+      "step": 1110
+    },
+    {
+      "epoch": 1.1941333333333333,
+      "grad_norm": 2.0834813633873823,
+      "learning_rate": 8.902227600533857e-06,
+      "loss": 0.2605243444442749,
+      "step": 1120
+    },
+    {
+      "epoch": 1.2048,
+      "grad_norm": 2.207036591702502,
+      "learning_rate": 8.880545371709749e-06,
+      "loss": 0.2487651824951172,
+      "step": 1130
+    },
+    {
+      "epoch": 1.2154666666666667,
+      "grad_norm": 2.2578329927979652,
+      "learning_rate": 8.858678063132649e-06,
+      "loss": 0.24483366012573243,
+      "step": 1140
+    },
+    {
+      "epoch": 1.2261333333333333,
+      "grad_norm": 2.10477492412832,
+      "learning_rate": 8.836626717747726e-06,
+      "loss": 0.24892668724060057,
+      "step": 1150
+    },
+    {
+      "epoch": 1.2368000000000001,
+      "grad_norm": 2.3387501935576642,
+      "learning_rate": 8.81439238727766e-06,
+      "loss": 0.24501357078552247,
+      "step": 1160
+    },
+    {
+      "epoch": 1.2474666666666667,
+      "grad_norm": 2.0333494452475285,
+      "learning_rate": 8.791976132172462e-06,
+      "loss": 0.24308676719665528,
+      "step": 1170
+    },
+    {
+      "epoch": 1.2581333333333333,
+      "grad_norm": 3.1826561387331753,
+      "learning_rate": 8.769379021558908e-06,
+      "loss": 0.25639915466308594,
+      "step": 1180
+    },
+    {
+      "epoch": 1.2688,
+      "grad_norm": 2.2808934157856444,
+      "learning_rate": 8.746602133189542e-06,
+      "loss": 0.2509211540222168,
+      "step": 1190
+    },
+    {
+      "epoch": 1.2794666666666665,
+      "grad_norm": 2.1930257037095187,
+      "learning_rate": 8.723646553391275e-06,
+      "loss": 0.2533258438110352,
+      "step": 1200
+    },
+    {
+      "epoch": 1.2901333333333334,
+      "grad_norm": 2.133767732753803,
+      "learning_rate": 8.70051337701358e-06,
+      "loss": 0.2539417266845703,
+      "step": 1210
+    },
+    {
+      "epoch": 1.3008,
+      "grad_norm": 2.239505180748716,
+      "learning_rate": 8.677203707376261e-06,
+      "loss": 0.2511929988861084,
+      "step": 1220
+    },
+    {
+      "epoch": 1.3114666666666666,
+      "grad_norm": 2.400110342470654,
+      "learning_rate": 8.65371865621684e-06,
+      "loss": 0.24864754676818848,
+      "step": 1230
+    },
+    {
+      "epoch": 1.3221333333333334,
+      "grad_norm": 2.074240653541272,
+      "learning_rate": 8.630059343637536e-06,
+      "loss": 0.25031208992004395,
+      "step": 1240
+    },
+    {
+      "epoch": 1.3328,
+      "grad_norm": 2.314210134369462,
+      "learning_rate": 8.606226898051832e-06,
+      "loss": 0.25528905391693113,
+      "step": 1250
+    },
+    {
+      "epoch": 1.3434666666666666,
+      "grad_norm": 2.266038410404579,
+      "learning_rate": 8.582222456130665e-06,
+      "loss": 0.2509054183959961,
+      "step": 1260
+    },
+    {
+      "epoch": 1.3541333333333334,
+      "grad_norm": 2.1656893509907746,
+      "learning_rate": 8.55804716274821e-06,
+      "loss": 0.2446660041809082,
+      "step": 1270
+    },
+    {
+      "epoch": 1.3648,
+      "grad_norm": 2.1062431094058067,
+      "learning_rate": 8.533702170927274e-06,
+      "loss": 0.24778637886047364,
+      "step": 1280
+    },
+    {
+      "epoch": 1.3754666666666666,
+      "grad_norm": 2.27440775778361,
+      "learning_rate": 8.509188641784313e-06,
+      "loss": 0.2531563758850098,
+      "step": 1290
+    },
+    {
+      "epoch": 1.3861333333333334,
+      "grad_norm": 2.0945350781678407,
+      "learning_rate": 8.484507744474043e-06,
+      "loss": 0.25156211853027344,
+      "step": 1300
+    },
+    {
+      "epoch": 1.3968,
+      "grad_norm": 2.344329246263725,
+      "learning_rate": 8.459660656133676e-06,
+      "loss": 0.25161404609680177,
+      "step": 1310
+    },
+    {
+      "epoch": 1.4074666666666666,
+      "grad_norm": 2.131637669904885,
+      "learning_rate": 8.434648561826791e-06,
+      "loss": 0.24704971313476562,
+      "step": 1320
+    },
+    {
+      "epoch": 1.4181333333333335,
+      "grad_norm": 2.084925140729102,
+      "learning_rate": 8.409472654486802e-06,
+      "loss": 0.2557060956954956,
+      "step": 1330
+    },
+    {
+      "epoch": 1.4288,
+      "grad_norm": 2.1914203263108405,
+      "learning_rate": 8.384134134860057e-06,
+      "loss": 0.2547765254974365,
+      "step": 1340
+    },
+    {
+      "epoch": 1.4394666666666667,
+      "grad_norm": 2.158051551717063,
+      "learning_rate": 8.358634211448595e-06,
+      "loss": 0.24558641910552978,
+      "step": 1350
+    },
+    {
+      "epoch": 1.4501333333333333,
+      "grad_norm": 2.2443418909572523,
+      "learning_rate": 8.332974100452472e-06,
+      "loss": 0.2537588357925415,
+      "step": 1360
+    },
+    {
+      "epoch": 1.4607999999999999,
+      "grad_norm": 2.0077376058555765,
+      "learning_rate": 8.307155025711783e-06,
+      "loss": 0.24577705860137938,
+      "step": 1370
+    },
+    {
+      "epoch": 1.4714666666666667,
+      "grad_norm": 2.065591654232565,
+      "learning_rate": 8.281178218648278e-06,
+      "loss": 0.25243277549743653,
+      "step": 1380
+    },
+    {
+      "epoch": 1.4821333333333333,
+      "grad_norm": 2.2061568477558744,
+      "learning_rate": 8.255044918206634e-06,
+      "loss": 0.2530541896820068,
+      "step": 1390
+    },
+    {
+      "epoch": 1.4928,
+      "grad_norm": 2.1819835453447842,
+      "learning_rate": 8.228756370795367e-06,
+      "loss": 0.2566534042358398,
+      "step": 1400
+    },
+    {
+      "epoch": 1.5034666666666667,
+      "grad_norm": 2.39971243280419,
+      "learning_rate": 8.20231383022738e-06,
+      "loss": 0.25281970500946044,
+      "step": 1410
+    },
+    {
+      "epoch": 1.5141333333333333,
+      "grad_norm": 2.1788579288513117,
+      "learning_rate": 8.175718557660168e-06,
+      "loss": 0.2553559303283691,
+      "step": 1420
+    },
+    {
+      "epoch": 1.5248,
+      "grad_norm": 1.9859089958121359,
+      "learning_rate": 8.148971821535662e-06,
+      "loss": 0.26154184341430664,
+      "step": 1430
+    },
+    {
+      "epoch": 1.5354666666666668,
+      "grad_norm": 2.2533326386546104,
+      "learning_rate": 8.122074897519743e-06,
+      "loss": 0.2603666543960571,
+      "step": 1440
+    },
+    {
+      "epoch": 1.5461333333333334,
+      "grad_norm": 2.3349705598874477,
+      "learning_rate": 8.095029068441388e-06,
+      "loss": 0.255629825592041,
+      "step": 1450
+    },
+    {
+      "epoch": 1.5568,
+      "grad_norm": 2.0149070325575544,
+      "learning_rate": 8.067835624231488e-06,
+      "loss": 0.2591118812561035,
+      "step": 1460
+    },
+    {
+      "epoch": 1.5674666666666668,
+      "grad_norm": 2.1532244460329966,
+      "learning_rate": 8.040495861861342e-06,
+      "loss": 0.25819978713989256,
+      "step": 1470
+    },
+    {
+      "epoch": 1.5781333333333334,
+      "grad_norm": 2.167458483789604,
+      "learning_rate": 8.013011085280769e-06,
+      "loss": 0.26111855506896975,
+      "step": 1480
+    },
+    {
+      "epoch": 1.5888,
+      "grad_norm": 2.3369281583312267,
+      "learning_rate": 7.985382605355945e-06,
+      "loss": 0.25337657928466795,
+      "step": 1490
+    },
+    {
+      "epoch": 1.5994666666666668,
+      "grad_norm": 2.146446091115663,
+      "learning_rate": 7.957611739806869e-06,
+      "loss": 0.2608675718307495,
+      "step": 1500
+    },
+    {
+      "epoch": 1.6101333333333332,
+      "grad_norm": 2.449035710299386,
+      "learning_rate": 7.929699813144515e-06,
+      "loss": 0.2538380861282349,
+      "step": 1510
+    },
+    {
+      "epoch": 1.6208,
+      "grad_norm": 2.022668293735054,
+      "learning_rate": 7.901648156607666e-06,
+      "loss": 0.2559537410736084,
+      "step": 1520
+    },
+    {
+      "epoch": 1.6314666666666666,
+      "grad_norm": 2.311749346530575,
+      "learning_rate": 7.873458108099414e-06,
+      "loss": 0.2684500217437744,
+      "step": 1530
+    },
+    {
+      "epoch": 1.6421333333333332,
+      "grad_norm": 2.3505705501308496,
+      "learning_rate": 7.845131012123364e-06,
+      "loss": 0.2538013458251953,
+      "step": 1540
+    },
+    {
+      "epoch": 1.6528,
+      "grad_norm": 2.2321743353517567,
+      "learning_rate": 7.816668219719483e-06,
+      "loss": 0.2512939691543579,
+      "step": 1550
+    },
+    {
+      "epoch": 1.6634666666666666,
+      "grad_norm": 2.328319256336755,
+      "learning_rate": 7.788071088399696e-06,
+      "loss": 0.2541865348815918,
+      "step": 1560
+    },
+    {
+      "epoch": 1.6741333333333333,
+      "grad_norm": 2.10587226339982,
+      "learning_rate": 7.75934098208311e-06,
+      "loss": 0.2653418302536011,
+      "step": 1570
+    },
+    {
+      "epoch": 1.6848,
+      "grad_norm": 2.090592027992588,
+      "learning_rate": 7.730479271030986e-06,
+      "loss": 0.24748015403747559,
+      "step": 1580
+    },
+    {
+      "epoch": 1.6954666666666667,
+      "grad_norm": 2.3497226188039275,
+      "learning_rate": 7.701487331781372e-06,
+      "loss": 0.2669647693634033,
+      "step": 1590
+    },
+    {
+      "epoch": 1.7061333333333333,
+      "grad_norm": 2.276620579505611,
+      "learning_rate": 7.672366547083454e-06,
+      "loss": 0.25644373893737793,
+      "step": 1600
+    },
+    {
+      "epoch": 1.7168,
+      "grad_norm": 2.1222637285476162,
+      "learning_rate": 7.643118305831605e-06,
+      "loss": 0.2490679979324341,
+      "step": 1610
+    },
+    {
+      "epoch": 1.7274666666666667,
+      "grad_norm": 2.229863229946843,
+      "learning_rate": 7.6137440029991505e-06,
+      "loss": 0.24733123779296876,
+      "step": 1620
+    },
+    {
+      "epoch": 1.7381333333333333,
+      "grad_norm": 2.2249450908491193,
+      "learning_rate": 7.584245039571822e-06,
+      "loss": 0.25619537830352784,
+      "step": 1630
+    },
+    {
+      "epoch": 1.7488000000000001,
+      "grad_norm": 2.2102876649687326,
+      "learning_rate": 7.554622822480953e-06,
+      "loss": 0.2534807205200195,
+      "step": 1640
+    },
+    {
+      "epoch": 1.7594666666666665,
+      "grad_norm": 2.2712866626459154,
+      "learning_rate": 7.524878764536368e-06,
+      "loss": 0.2548924446105957,
+      "step": 1650
+    },
+    {
+      "epoch": 1.7701333333333333,
+      "grad_norm": 2.22733165752973,
+      "learning_rate": 7.495014284358999e-06,
+      "loss": 0.24510478973388672,
+      "step": 1660
+    },
+    {
+      "epoch": 1.7808000000000002,
+      "grad_norm": 2.099299914243324,
+      "learning_rate": 7.465030806313227e-06,
+      "loss": 0.2522265434265137,
+      "step": 1670
+    },
+    {
+      "epoch": 1.7914666666666665,
+      "grad_norm": 2.197727121468574,
+      "learning_rate": 7.4349297604389555e-06,
+      "loss": 0.26017649173736573,
+      "step": 1680
+    },
+    {
+      "epoch": 1.8021333333333334,
+      "grad_norm": 2.2740805256970096,
+      "learning_rate": 7.40471258238339e-06,
+      "loss": 0.2631208896636963,
+      "step": 1690
+    },
+    {
+      "epoch": 1.8128,
+      "grad_norm": 2.235596970175843,
+      "learning_rate": 7.374380713332581e-06,
+      "loss": 0.2545802354812622,
+      "step": 1700
+    },
+    {
+      "epoch": 1.8234666666666666,
+      "grad_norm": 2.1858944714517463,
+      "learning_rate": 7.343935599942681e-06,
+      "loss": 0.25033328533172605,
+      "step": 1710
+    },
+    {
+      "epoch": 1.8341333333333334,
+      "grad_norm": 2.2088385931044066,
+      "learning_rate": 7.313378694270943e-06,
+      "loss": 0.2528871536254883,
+      "step": 1720
+    },
+    {
+      "epoch": 1.8448,
+      "grad_norm": 2.162468266561391,
+      "learning_rate": 7.282711453706477e-06,
+      "loss": 0.25245530605316163,
+      "step": 1730
+    },
+    {
+      "epoch": 1.8554666666666666,
+      "grad_norm": 2.2054248406128294,
+      "learning_rate": 7.251935340900731e-06,
+      "loss": 0.2531111240386963,
+      "step": 1740
+    },
+    {
+      "epoch": 1.8661333333333334,
+      "grad_norm": 2.674099478078809,
+      "learning_rate": 7.221051823697734e-06,
+      "loss": 0.25325212478637693,
+      "step": 1750
+    },
+    {
+      "epoch": 1.8768,
+      "grad_norm": 2.2247776209641494,
+      "learning_rate": 7.190062375064089e-06,
+      "loss": 0.24999377727508545,
+      "step": 1760
+    },
+    {
+      "epoch": 1.8874666666666666,
+      "grad_norm": 2.200346202814715,
+      "learning_rate": 7.15896847301872e-06,
+      "loss": 0.25710105895996094,
+      "step": 1770
+    },
+    {
+      "epoch": 1.8981333333333335,
+      "grad_norm": 2.6602196093964783,
+      "learning_rate": 7.127771600562385e-06,
+      "loss": 0.24712116718292237,
+      "step": 1780
+    },
+    {
+      "epoch": 1.9088,
+      "grad_norm": 2.0979266321264083,
+      "learning_rate": 7.0964732456069304e-06,
+      "loss": 0.2447052478790283,
+      "step": 1790
+    },
+    {
+      "epoch": 1.9194666666666667,
+      "grad_norm": 2.0469909236188264,
+      "learning_rate": 7.065074900904344e-06,
+      "loss": 0.2608736515045166,
+      "step": 1800
+    },
+    {
+      "epoch": 1.9301333333333335,
+      "grad_norm": 2.1567998095922456,
+      "learning_rate": 7.033578063975546e-06,
+      "loss": 0.2497999668121338,
+      "step": 1810
+    },
+    {
+      "epoch": 1.9407999999999999,
+      "grad_norm": 2.10629682011714,
+      "learning_rate": 7.0019842370389715e-06,
+      "loss": 0.24511592388153075,
+      "step": 1820
+    },
+    {
+      "epoch": 1.9514666666666667,
+      "grad_norm": 2.285009344275457,
+      "learning_rate": 6.9702949269389254e-06,
+      "loss": 0.2560229778289795,
+      "step": 1830
+    },
+    {
+      "epoch": 1.9621333333333333,
+      "grad_norm": 2.1580134702270755,
+      "learning_rate": 6.938511645073709e-06,
+      "loss": 0.24839389324188232,
+      "step": 1840
+    },
+    {
+      "epoch": 1.9727999999999999,
+      "grad_norm": 2.059285536180603,
+      "learning_rate": 6.906635907323538e-06,
+      "loss": 0.24895498752593995,
+      "step": 1850
+    },
+    {
+      "epoch": 1.9834666666666667,
+      "grad_norm": 2.105134983270051,
+      "learning_rate": 6.874669233978245e-06,
+      "loss": 0.24931104183197023,
+      "step": 1860
+    },
+    {
+      "epoch": 1.9941333333333333,
+      "grad_norm": 2.1881330324208808,
+      "learning_rate": 6.842613149664766e-06,
+      "loss": 0.2481091022491455,
+      "step": 1870
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 4690,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 5,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 4.935356496604037e+17,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-1876/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d7601908377509110784aaf5570cd72f8058db6fa691374b2de503eee9ab4c4c
+size 7441

checkpoint-1876/zero_to_fp32.py

@@ -0,0 +1,760 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example:
+#   python zero_to_fp32.py . output_dir/
+#   or
+#   python zero_to_fp32.py . output_dir/ --safe_serialization
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+import gc
+import json
+import numpy as np
+from tqdm import tqdm
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device, weights_only=False)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+    total_files = len(files)
+    state_dicts = []
+    for f in tqdm(files, desc='Loading checkpoint shards'):
+        state_dict = torch.load(f, map_location=device, mmap=True, weights_only=False)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if ZERO_STAGE not in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+class GatheredTensor:
+    """
+    A pseudo tensor that collects partitioned weights.
+    It is more memory efficient when there are multiple groups.
+    """
+
+    def __init__(self, flat_groups, flat_groups_offset, offset, partitioned_numel, shape):
+        self.flat_groups = flat_groups
+        self.flat_groups_offset = flat_groups_offset
+        self.offset = offset
+        self.partitioned_numel = partitioned_numel
+        self.shape = shape
+        self.dtype = self.flat_groups[0][0].dtype
+
+    def contiguous(self):
+        """
+        Merge partitioned weights from flat_groups into a single tensor.
+        """
+        end_idx = self.offset + self.partitioned_numel
+        world_size = len(self.flat_groups)
+        pad_flat_param_chunks = []
+
+        for rank_i in range(world_size):
+            # for each rank, we need to collect weights from related group/groups
+            flat_groups_at_rank_i = self.flat_groups[rank_i]
+            start_group_id = None
+            end_group_id = None
+            for group_id in range(len(self.flat_groups_offset)):
+                if self.flat_groups_offset[group_id] <= self.offset < self.flat_groups_offset[group_id + 1]:
+                    start_group_id = group_id
+                if self.flat_groups_offset[group_id] < end_idx <= self.flat_groups_offset[group_id + 1]:
+                    end_group_id = group_id
+                    break
+            # collect weights from related group/groups
+            for group_id in range(start_group_id, end_group_id + 1):
+                flat_tensor = flat_groups_at_rank_i[group_id]
+                start_offset = self.offset - self.flat_groups_offset[group_id]
+                end_offset = min(end_idx, self.flat_groups_offset[group_id + 1]) - self.flat_groups_offset[group_id]
+                pad_flat_param_chunks.append(flat_tensor[start_offset:end_offset])
+
+        # collect weights from all ranks
+        pad_flat_param = torch.cat(pad_flat_param_chunks, dim=0)
+        param = pad_flat_param[:self.shape.numel()].view(self.shape).contiguous()
+        return param
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = sum([flat_group.numel() for flat_group in fp32_flat_groups[0]]) * world_size
+
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    flat_groups_offset = [0] + list(np.cumsum([flat_tensor.numel() for flat_tensor in fp32_flat_groups[0]]))
+    for name, shape in tqdm(param_shapes.items(), desc='Gathering sharded weights'):
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # memory efficient tensor
+        tensor = GatheredTensor(fp32_flat_groups, flat_groups_offset, offset, partitioned_numel, shape)
+        state_dict[name] = tensor
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def to_torch_tensor(state_dict, return_empty_tensor=False):
+    """
+    Convert state_dict of GatheredTensor to torch tensor
+    """
+    torch_state_dict = {}
+    converted_tensors = {}
+    for name, tensor in state_dict.items():
+        tensor_id = id(tensor)
+        if tensor_id in converted_tensors:  # shared tensors
+            shared_tensor = torch_state_dict[converted_tensors[tensor_id]]
+            torch_state_dict[name] = shared_tensor
+        else:
+            converted_tensors[tensor_id] = name
+            if return_empty_tensor:
+                torch_state_dict[name] = torch.empty(tensor.shape, dtype=tensor.dtype)
+            else:
+                torch_state_dict[name] = tensor.contiguous()
+    return torch_state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir,
+                                             tag=None,
+                                             exclude_frozen_parameters=False,
+                                             lazy_mode=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+        - ``lazy_mode``: get state_dict in lazy mode. It returns a dict of pesduo tensor instead of torch tensor, which is more memory efficient.
+          Convert the pesduo tensor to torch tensor by ``.contiguous()``
+
+    Returns:
+        - pytorch ``state_dict``
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    Note: the above usage may not work if your application doesn't have sufficient free CPU memory.
+    You may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint. Or you can load state_dict in lazy mode ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, lazy_mode=True) # not on cpu
+        for name, lazy_tensor in state_dict.item():
+            tensor = lazy_tensor.contiguous()  # to cpu
+            print(name, tensor)
+            # del tensor to release memory if it no longer in use
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    state_dict = _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+    if lazy_mode:
+        return state_dict
+    else:
+        return to_torch_tensor(state_dict)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir,
+                                               output_dir,
+                                               max_shard_size="5GB",
+                                               safe_serialization=False,
+                                               tag=None,
+                                               exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_dir``: directory to the pytorch fp32 state_dict output files
+        - ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB
+        - ``safe_serialization``:  whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    # Dependency pre-check
+    if safe_serialization:
+        try:
+            from safetensors.torch import save_file
+        except ImportError:
+            print('If you want to use `safe_serialization`, please `pip install safetensors`')
+            raise
+    if max_shard_size is not None:
+        try:
+            from huggingface_hub import split_torch_state_dict_into_shards
+        except ImportError:
+            print('If you want to use `max_shard_size`, please `pip install huggingface_hub`')
+            raise
+
+    # Convert zero checkpoint to state_dict
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir,
+                                                          tag,
+                                                          exclude_frozen_parameters,
+                                                          lazy_mode=True)
+
+    # Shard the model if it is too big.
+    weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin"
+    if max_shard_size is not None:
+        filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors")
+        # an memory-efficient approach for sharding
+        empty_state_dict = to_torch_tensor(state_dict, return_empty_tensor=True)
+        state_dict_split = split_torch_state_dict_into_shards(empty_state_dict,
+                                                              filename_pattern=filename_pattern,
+                                                              max_shard_size=max_shard_size)
+    else:
+        from collections import namedtuple
+        StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"])
+        state_dict_split = StateDictSplit(is_sharded=False,
+                                          filename_to_tensors={weights_name: list(state_dict.keys())})
+
+    # Save the model by shard
+    os.makedirs(output_dir, exist_ok=True)
+    filename_to_tensors = state_dict_split.filename_to_tensors.items()
+    for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
+        shard_state_dict = {tensor_name: state_dict[tensor_name] for tensor_name in tensors}
+        shard_state_dict = to_torch_tensor(shard_state_dict)
+        output_path = os.path.join(output_dir, shard_file)
+        if safe_serialization:
+            save_file(shard_state_dict, output_path, metadata={"format": "pt"})
+        else:
+            torch.save(shard_state_dict, output_path)
+        # release the memory of current shard
+        for tensor_name in list(shard_state_dict.keys()):
+            del state_dict[tensor_name]
+            del shard_state_dict[tensor_name]
+        del shard_state_dict
+        gc.collect()
+
+    # Save index if sharded
+    if state_dict_split.is_sharded:
+        index = {
+            "metadata": state_dict_split.metadata,
+            "weight_map": state_dict_split.tensor_to_filename,
+        }
+        save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json"
+        save_index_file = os.path.join(output_dir, save_index_file)
+        with open(save_index_file, "w", encoding="utf-8") as f:
+            content = json.dumps(index, indent=2, sort_keys=True) + "\n"
+            f.write(content)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info("Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info("Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument("output_dir",
+                        type=str,
+                        help="directory to the pytorch fp32 state_dict output files"
+                        "(e.g. path/checkpoint-12-output/)")
+    parser.add_argument(
+        "--max_shard_size",
+        type=str,
+        default="5GB",
+        help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size"
+        "lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`"
+        "We default it to 5GB in order for models to be able to run easily on free-tier google colab instances"
+        "without CPU OOM issues.")
+    parser.add_argument(
+        "--safe_serialization",
+        default=False,
+        action='store_true',
+        help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_dir,
+                                               max_shard_size=args.max_shard_size,
+                                               safe_serialization=args.safe_serialization,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)

checkpoint-2814/chat_template.jinja

@@ -0,0 +1,154 @@
+{%- set image_count = namespace(value=0) %}
+{%- set video_count = namespace(value=0) %}
+{%- macro render_content(content, do_vision_count, is_system_content=false) %}
+    {%- if content is string %}
+        {{- content }}
+    {%- elif content is iterable and content is not mapping %}
+        {%- for item in content %}
+            {%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
+                {%- if is_system_content %}
+                    {{- raise_exception('System message cannot contain images.') }}
+                {%- endif %}
+                {%- if do_vision_count %}
+                    {%- set image_count.value = image_count.value + 1 %}
+                {%- endif %}
+                {%- if add_vision_id %}
+                    {{- 'Picture ' ~ image_count.value ~ ': ' }}
+                {%- endif %}
+                {{- '<|vision_start|><|image_pad|><|vision_end|>' }}
+            {%- elif 'video' in item or item.type == 'video' %}
+                {%- if is_system_content %}
+                    {{- raise_exception('System message cannot contain videos.') }}
+                {%- endif %}
+                {%- if do_vision_count %}
+                    {%- set video_count.value = video_count.value + 1 %}
+                {%- endif %}
+                {%- if add_vision_id %}
+                    {{- 'Video ' ~ video_count.value ~ ': ' }}
+                {%- endif %}
+                {{- '<|vision_start|><|video_pad|><|vision_end|>' }}
+            {%- elif 'text' in item %}
+                {{- item.text }}
+            {%- else %}
+                {{- raise_exception('Unexpected item type in content.') }}
+            {%- endif %}
+        {%- endfor %}
+    {%- elif content is none or content is undefined %}
+        {{- '' }}
+    {%- else %}
+        {{- raise_exception('Unexpected content type.') }}
+    {%- endif %}
+{%- endmacro %}
+{%- if not messages %}
+    {{- raise_exception('No messages provided.') }}
+{%- endif %}
+{%- if tools and tools is iterable and tools is not mapping %}
+    {{- '<|im_start|>system\n' }}
+    {{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>" }}
+    {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <tool_call></tool_call> XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n</IMPORTANT>' }}
+    {%- if messages[0].role == 'system' %}
+        {%- set content = render_content(messages[0].content, false, true)|trim %}
+        {%- if content %}
+            {{- '\n\n' + content }}
+        {%- endif %}
+    {%- endif %}
+    {{- '<|im_end|>\n' }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {%- set content = render_content(messages[0].content, false, true)|trim %}
+        {{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+    {%- set index = (messages|length - 1) - loop.index0 %}
+    {%- if ns.multi_step_tool and message.role == "user" %}
+        {%- set content = render_content(message.content, false)|trim %}
+        {%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
+            {%- set ns.multi_step_tool = false %}
+            {%- set ns.last_query_index = index %}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if ns.multi_step_tool %}
+    {{- raise_exception('No user query found in messages.') }}
+{%- endif %}
+{%- for message in messages %}
+    {%- set content = render_content(message.content, true)|trim %}
+    {%- if message.role == "system" %}
+        {%- if not loop.first %}
+            {{- raise_exception('System message must be at the beginning.') }}
+        {%- endif %}
+    {%- elif message.role == "user" %}
+        {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is string %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in content %}
+                {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+                {%- set content = content.split('</think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- set reasoning_content = reasoning_content|trim %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\n\n' + content }}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if tool_call.function is defined %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {%- if loop.first %}
+                    {%- if content|trim %}
+                        {{- '\n\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
+                    {%- else %}
+                        {{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
+                    {%- endif %}
+                {%- else %}
+                    {{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
+                {%- endif %}
+                {%- if tool_call.arguments is defined %}
+                    {%- for args_name, args_value in tool_call.arguments|items %}
+                        {{- '<parameter=' + args_name + '>\n' }}
+                        {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %}
+                        {{- args_value }}
+                        {{- '\n</parameter>\n' }}
+                    {%- endfor %}
+                {%- endif %}
+                {{- '</function>\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.previtem and loop.previtem.role != "tool" %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- content }}
+        {{- '\n</tool_response>' }}
+        {%- if not loop.last and loop.nextitem.role != "tool" %}
+            {{- '<|im_end|>\n' }}
+        {%- elif loop.last %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- else %}
+        {{- raise_exception('Unexpected message role.') }}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+    {%- if enable_thinking is defined and enable_thinking is false %}
+        {{- '<think>\n\n</think>\n\n' }}
+    {%- else %}
+        {{- '<think>\n' }}
+    {%- endif %}
+{%- endif %}

checkpoint-2814/config.json

@@ -0,0 +1,113 @@
+{
+  "architectures": [
+    "Qwen3_5ForConditionalGeneration"
+  ],
+  "dtype": "bfloat16",
+  "eos_token_id": 248046,
+  "hidden_size": 4096,
+  "image_token_id": 248056,
+  "model_type": "qwen3_5",
+  "pad_token_id": 248044,
+  "text_config": {
+    "attention_bias": false,
+    "attention_dropout": 0.0,
+    "attn_output_gate": true,
+    "bos_token_id": null,
+    "dtype": "bfloat16",
+    "eos_token_id": 248044,
+    "full_attention_interval": 4,
+    "head_dim": 256,
+    "hidden_act": "silu",
+    "hidden_size": 4096,
+    "initializer_range": 0.02,
+    "intermediate_size": 12288,
+    "layer_types": [
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention",
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention",
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention",
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention",
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention",
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention",
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention",
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention"
+    ],
+    "linear_conv_kernel_dim": 4,
+    "linear_key_head_dim": 128,
+    "linear_num_key_heads": 16,
+    "linear_num_value_heads": 32,
+    "linear_value_head_dim": 128,
+    "mamba_ssm_dtype": "float32",
+    "max_position_embeddings": 262144,
+    "mlp_only_layers": [],
+    "model_type": "qwen3_5_text",
+    "mtp_num_hidden_layers": 1,
+    "mtp_use_dedicated_embeddings": false,
+    "num_attention_heads": 16,
+    "num_hidden_layers": 32,
+    "num_key_value_heads": 4,
+    "pad_token_id": null,
+    "partial_rotary_factor": 0.25,
+    "rms_norm_eps": 1e-06,
+    "rope_parameters": {
+      "mrope_interleaved": true,
+      "mrope_section": [
+        11,
+        11,
+        10
+      ],
+      "partial_rotary_factor": 0.25,
+      "rope_theta": 10000000,
+      "rope_type": "default"
+    },
+    "tie_word_embeddings": false,
+    "use_cache": false,
+    "vocab_size": 248320
+  },
+  "tie_word_embeddings": false,
+  "transformers_version": "5.6.0",
+  "use_cache": false,
+  "video_token_id": 248057,
+  "vision_config": {
+    "deepstack_visual_indexes": [],
+    "depth": 27,
+    "dtype": "bfloat16",
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 1152,
+    "in_channels": 3,
+    "initializer_range": 0.02,
+    "intermediate_size": 4304,
+    "model_type": "qwen3_5_vision",
+    "num_heads": 16,
+    "num_position_embeddings": 2304,
+    "out_hidden_size": 4096,
+    "patch_size": 16,
+    "spatial_merge_size": 2,
+    "temporal_patch_size": 2
+  },
+  "vision_end_token_id": 248054,
+  "vision_start_token_id": 248053
+}

checkpoint-2814/generation_config.json

@@ -0,0 +1,10 @@
+{
+  "_from_model_config": true,
+  "eos_token_id": [
+    248046,
+    248044
+  ],
+  "pad_token_id": 248044,
+  "transformers_version": "5.6.0",
+  "use_cache": true
+}

checkpoint-2814/global_step2814/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3f4c563ace3bd52defcd0322f2f1e18b7e601162d785cc4168cad9634fd81610
+size 13430710881

checkpoint-2814/global_step2814/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77ee383c8c79319610c455d709d1b731285da847a036dbe3ee9d384307c61837
+size 13430710881

checkpoint-2814/global_step2814/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:19abab949f17739e4eb968f44d149893e0ce1adb6bfc4808fafbabbbf114e5c8
+size 13430710881

checkpoint-2814/global_step2814/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8705c21e99bfb9993971500af26a9535c6ff0b0b72bd1baa68190893c8c300b9
+size 13430710881

checkpoint-2814/global_step2814/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aeda2bc1c6679b09a78b2ffe7ade11792d1b6401217a3f92fc84ea3463835913
+size 13430710881

checkpoint-2814/global_step2814/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4084ef5834ef0d92546b11a03c59e0d3b6ec6635fc0aca8866e8569191013000
+size 13430710881

checkpoint-2814/global_step2814/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e2b062293ee7405a271e47e54a3ad4f6128b8a6b5a4f5b5bab8b352a518010a2
+size 13430710881