Upload folder using huggingface_hub

.gitattributes

@@ -44,3 +44,4 @@ checkpoints/qwen3/qwen3_p1621m_t1_16_suite-owt_scaling_v3_family-qwen3_geometry_
 checkpoints/qwen3/qwen3_p223m_t1_128_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt1_20260211_204923/checkpoint-251/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 checkpoints/qwen3/qwen3_p047m_t1_32_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt1_20260211_235341/checkpoint-768/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 checkpoints/qwen3/qwen3_p1083m_t1_16_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt1_20260210_122411/checkpoint-768/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/tokenizer.json filter=lfs diff=lfs merge=lfs -text

checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/added_tokens.json

@@ -0,0 +1,28 @@
+{
+  "</think>": 151668,
+  "</tool_call>": 151658,
+  "</tool_response>": 151666,
+  "<think>": 151667,
+  "<tool_call>": 151657,
+  "<tool_response>": 151665,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/config.json

@@ -0,0 +1,57 @@
+{
+  "architectures": [
+    "Qwen3ForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "dtype": "bfloat16",
+  "eos_token_id": 151645,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 1536,
+  "initializer_range": 0.02,
+  "intermediate_size": 4608,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 40960,
+  "max_window_layers": 24,
+  "model_type": "qwen3",
+  "num_attention_heads": 24,
+  "num_hidden_layers": 24,
+  "num_key_value_heads": 12,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000,
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.51.3",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}

checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/generation_config.json

@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "transformers_version": "4.51.3",
+  "use_cache": false
+}

checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/global_step8192/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:ae4c63ffb3ff09b0b155479f717b42245e3b525170d12bcc81ca8402c9a151ec
+size 3248411504

checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/global_step8192/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5e93cf221e1187582a0b0b41ff6833f85088cc506da6496825b1b09f44804d00
+size 3248416112

checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/global_step8192/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:819ead047105bf193a0b3315fb436858301b0c8167b91873b9384c6bc8070469
+size 3248416048

checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/global_step8192/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:a2147e05313806cf205f4f0f8155f27ab21b451806e8d858c99cc7337fb18ab3
+size 3248416112

checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/global_step8192/mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a19f89b00fe63d524a07a7634499da4903d37d045541831a4dc43eee0d4c675d
+size 2165680184

checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/latest

@@ -0,0 +1 @@
+global_step8192

checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:88470a67430b03bcaf110104464fa439d4677b8d70fc02312e1779d20644bb9f
+size 2632381440

checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:543ef05f530d40ee20b8d626b07a69b86597aca643e48897571062f973efe84f
+size 15024

checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7a23f732e43838ce0398d2636885ac16badbb9bcbc04d1406069ba3027bc5ae0
+size 15024

checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e10cce960e7068b051c05e35ed6160656be9091c63f13796ac2ed7e9c84e5a72
+size 15024

checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b6f6049e212b1df5cefc5d834afcd8cc052c73f1457449e9fe8a38d514f54078
+size 15024

checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:334fb076f7d71bb621987f565fce1446b156844c3df0ecee01029beb434ae082
+size 1064

checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/special_tokens_map.json

@@ -0,0 +1,31 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4
+size 11422654

checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/tokenizer_config.json

@@ -0,0 +1,241 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151665": {
+      "content": "<tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151666": {
+      "content": "</tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151667": {
+      "content": "<think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151668": {
+      "content": "</think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0].role == 'system' %}\n        {{- messages[0].content + '\\n\\n' }}\n    {%- endif %}\n    {{- \"# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0].role == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0].content + '<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}\n{%- for message in messages[::-1] %}\n    {%- set index = (messages|length - 1) - loop.index0 %}\n    {%- if ns.multi_step_tool and message.role == \"user\" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}\n        {%- set ns.multi_step_tool = false %}\n        {%- set ns.last_query_index = index %}\n    {%- endif %}\n{%- endfor %}\n{%- for message in messages %}\n    {%- if message.content is string %}\n        {%- set content = message.content %}\n    {%- else %}\n        {%- set content = '' %}\n    {%- endif %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) %}\n        {{- '<|im_start|>' + message.role + '\\n' + content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {%- set reasoning_content = '' %}\n        {%- if message.reasoning_content is string %}\n            {%- set reasoning_content = message.reasoning_content %}\n        {%- else %}\n            {%- if '</think>' in content %}\n                {%- set reasoning_content = content.split('</think>')[0].rstrip('\\n').split('<think>')[-1].lstrip('\\n') %}\n                {%- set content = content.split('</think>')[-1].lstrip('\\n') %}\n            {%- endif %}\n        {%- endif %}\n        {%- if loop.index0 > ns.last_query_index %}\n            {%- if loop.last or (not loop.last and reasoning_content) %}\n                {{- '<|im_start|>' + message.role + '\\n<think>\\n' + reasoning_content.strip('\\n') + '\\n</think>\\n\\n' + content.lstrip('\\n') }}\n            {%- else %}\n                {{- '<|im_start|>' + message.role + '\\n' + content }}\n            {%- endif %}\n        {%- else %}\n            {{- '<|im_start|>' + message.role + '\\n' + content }}\n        {%- endif %}\n        {%- if message.tool_calls %}\n            {%- for tool_call in message.tool_calls %}\n                {%- if (loop.first and content) or (not loop.first) %}\n                    {{- '\\n' }}\n                {%- endif %}\n                {%- if tool_call.function %}\n                    {%- set tool_call = tool_call.function %}\n                {%- endif %}\n                {{- '<tool_call>\\n{\"name\": \"' }}\n                {{- tool_call.name }}\n                {{- '\", \"arguments\": ' }}\n                {%- if tool_call.arguments is string %}\n                    {{- tool_call.arguments }}\n                {%- else %}\n                    {{- tool_call.arguments | tojson }}\n                {%- endif %}\n                {{- '}\\n</tool_call>' }}\n            {%- endfor %}\n        {%- endif %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if loop.first or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n    {%- if enable_thinking is defined and enable_thinking is false %}\n        {{- '<think>\\n\\n</think>\\n\\n' }}\n    {%- endif %}\n{%- endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/trainer_state.json

@@ -0,0 +1,665 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.24644727985168638,
+  "eval_steps": 1024,
+  "global_step": 8192,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.003008389646627031,
+      "grad_norm": 1.8081544637680054,
+      "learning_rate": 3.08411214953271e-05,
+      "loss": 9.7352,
+      "step": 100
+    },
+    {
+      "epoch": 0.006016779293254062,
+      "grad_norm": 1.0897260904312134,
+      "learning_rate": 6.199376947040498e-05,
+      "loss": 7.0092,
+      "step": 200
+    },
+    {
+      "epoch": 0.009025168939881093,
+      "grad_norm": 1.0756351947784424,
+      "learning_rate": 9.314641744548286e-05,
+      "loss": 6.2106,
+      "step": 300
+    },
+    {
+      "epoch": 0.012033558586508125,
+      "grad_norm": 1.325608730316162,
+      "learning_rate": 9.999390218499588e-05,
+      "loss": 5.7481,
+      "step": 400
+    },
+    {
+      "epoch": 0.015041948233135156,
+      "grad_norm": 0.5563725233078003,
+      "learning_rate": 9.996824677076817e-05,
+      "loss": 5.4152,
+      "step": 500
+    },
+    {
+      "epoch": 0.018050337879762186,
+      "grad_norm": 0.6096261143684387,
+      "learning_rate": 9.992255893455041e-05,
+      "loss": 5.17,
+      "step": 600
+    },
+    {
+      "epoch": 0.021058727526389216,
+      "grad_norm": 0.7971847057342529,
+      "learning_rate": 9.9856856992735e-05,
+      "loss": 4.9735,
+      "step": 700
+    },
+    {
+      "epoch": 0.02406711717301625,
+      "grad_norm": 0.49311694502830505,
+      "learning_rate": 9.977116728543007e-05,
+      "loss": 4.8053,
+      "step": 800
+    },
+    {
+      "epoch": 0.02707550681964328,
+      "grad_norm": 0.6969898343086243,
+      "learning_rate": 9.966552416589973e-05,
+      "loss": 4.6446,
+      "step": 900
+    },
+    {
+      "epoch": 0.030083896466270313,
+      "grad_norm": 0.5000762939453125,
+      "learning_rate": 9.953996998679162e-05,
+      "loss": 4.5142,
+      "step": 1000
+    },
+    {
+      "epoch": 0.030805909981460797,
+      "eval_loss": 4.429453372955322,
+      "eval_runtime": 28.6882,
+      "eval_samples_per_second": 147.726,
+      "eval_steps_per_second": 4.636,
+      "step": 1024
+    },
+    {
+      "epoch": 0.03309228611289734,
+      "grad_norm": 0.4743868112564087,
+      "learning_rate": 9.939455508315775e-05,
+      "loss": 4.4023,
+      "step": 1100
+    },
+    {
+      "epoch": 0.03610067575952437,
+      "grad_norm": 0.5583599805831909,
+      "learning_rate": 9.92293377522749e-05,
+      "loss": 4.3164,
+      "step": 1200
+    },
+    {
+      "epoch": 0.0391090654061514,
+      "grad_norm": 0.442892462015152,
+      "learning_rate": 9.904438423027316e-05,
+      "loss": 4.2406,
+      "step": 1300
+    },
+    {
+      "epoch": 0.04211745505277843,
+      "grad_norm": 0.3826092779636383,
+      "learning_rate": 9.883976866558158e-05,
+      "loss": 4.1792,
+      "step": 1400
+    },
+    {
+      "epoch": 0.04512584469940547,
+      "grad_norm": 0.42428717017173767,
+      "learning_rate": 9.86155730892019e-05,
+      "loss": 4.1225,
+      "step": 1500
+    },
+    {
+      "epoch": 0.0481342343460325,
+      "grad_norm": 0.394997239112854,
+      "learning_rate": 9.837188738182205e-05,
+      "loss": 4.07,
+      "step": 1600
+    },
+    {
+      "epoch": 0.05114262399265953,
+      "grad_norm": 0.36051058769226074,
+      "learning_rate": 9.81088092377827e-05,
+      "loss": 4.0229,
+      "step": 1700
+    },
+    {
+      "epoch": 0.05415101363928656,
+      "grad_norm": 0.3929230272769928,
+      "learning_rate": 9.782644412591135e-05,
+      "loss": 3.9876,
+      "step": 1800
+    },
+    {
+      "epoch": 0.05715940328591359,
+      "grad_norm": 0.39373254776000977,
+      "learning_rate": 9.752490524723946e-05,
+      "loss": 3.9443,
+      "step": 1900
+    },
+    {
+      "epoch": 0.060167792932540626,
+      "grad_norm": 0.3637371361255646,
+      "learning_rate": 9.720431348961991e-05,
+      "loss": 3.9091,
+      "step": 2000
+    },
+    {
+      "epoch": 0.061611819962921595,
+      "eval_loss": 3.878706216812134,
+      "eval_runtime": 27.5639,
+      "eval_samples_per_second": 153.752,
+      "eval_steps_per_second": 4.825,
+      "step": 2048
+    },
+    {
+      "epoch": 0.06317618257916766,
+      "grad_norm": 0.3592907190322876,
+      "learning_rate": 9.686479737926266e-05,
+      "loss": 3.8722,
+      "step": 2100
+    },
+    {
+      "epoch": 0.06618457222579469,
+      "grad_norm": 0.34135356545448303,
+      "learning_rate": 9.650649302920819e-05,
+      "loss": 3.851,
+      "step": 2200
+    },
+    {
+      "epoch": 0.06919296187242172,
+      "grad_norm": 0.34740984439849854,
+      "learning_rate": 9.612954408475934e-05,
+      "loss": 3.8182,
+      "step": 2300
+    },
+    {
+      "epoch": 0.07220135151904875,
+      "grad_norm": 0.3434409201145172,
+      "learning_rate": 9.573410166589356e-05,
+      "loss": 3.8004,
+      "step": 2400
+    },
+    {
+      "epoch": 0.07520974116567578,
+      "grad_norm": 0.30977728962898254,
+      "learning_rate": 9.532032430667834e-05,
+      "loss": 3.7671,
+      "step": 2500
+    },
+    {
+      "epoch": 0.0782181308123028,
+      "grad_norm": 0.3274868428707123,
+      "learning_rate": 9.488837789171455e-05,
+      "loss": 3.7498,
+      "step": 2600
+    },
+    {
+      "epoch": 0.08122652045892984,
+      "grad_norm": 0.3402297794818878,
+      "learning_rate": 9.443843558963266e-05,
+      "loss": 3.7235,
+      "step": 2700
+    },
+    {
+      "epoch": 0.08423491010555686,
+      "grad_norm": 0.33715498447418213,
+      "learning_rate": 9.397067778366902e-05,
+      "loss": 3.7161,
+      "step": 2800
+    },
+    {
+      "epoch": 0.08724329975218391,
+      "grad_norm": 0.32197704911231995,
+      "learning_rate": 9.348529199934959e-05,
+      "loss": 3.6971,
+      "step": 2900
+    },
+    {
+      "epoch": 0.09025168939881094,
+      "grad_norm": 0.31326088309288025,
+      "learning_rate": 9.298247282931048e-05,
+      "loss": 3.6674,
+      "step": 3000
+    },
+    {
+      "epoch": 0.0924177299443824,
+      "eval_loss": 3.6544880867004395,
+      "eval_runtime": 27.8043,
+      "eval_samples_per_second": 152.422,
+      "eval_steps_per_second": 4.783,
+      "step": 3072
+    },
+    {
+      "epoch": 0.09326007904543797,
+      "grad_norm": 0.3117671608924866,
+      "learning_rate": 9.246242185528508e-05,
+      "loss": 3.6553,
+      "step": 3100
+    },
+    {
+      "epoch": 0.096268468692065,
+      "grad_norm": 0.31069496273994446,
+      "learning_rate": 9.192534756728936e-05,
+      "loss": 3.6431,
+      "step": 3200
+    },
+    {
+      "epoch": 0.09927685833869203,
+      "grad_norm": 0.31252533197402954,
+      "learning_rate": 9.137146528003762e-05,
+      "loss": 3.6309,
+      "step": 3300
+    },
+    {
+      "epoch": 0.10228524798531906,
+      "grad_norm": 0.29626986384391785,
+      "learning_rate": 9.080099704662214e-05,
+      "loss": 3.6096,
+      "step": 3400
+    },
+    {
+      "epoch": 0.10529363763194609,
+      "grad_norm": 0.3036186993122101,
+      "learning_rate": 9.021417156949143e-05,
+      "loss": 3.6025,
+      "step": 3500
+    },
+    {
+      "epoch": 0.10830202727857312,
+      "grad_norm": 0.3038487136363983,
+      "learning_rate": 8.961122410876268e-05,
+      "loss": 3.5854,
+      "step": 3600
+    },
+    {
+      "epoch": 0.11131041692520015,
+      "grad_norm": 0.2864755094051361,
+      "learning_rate": 8.899239638790535e-05,
+      "loss": 3.5725,
+      "step": 3700
+    },
+    {
+      "epoch": 0.11431880657182718,
+      "grad_norm": 0.2870684862136841,
+      "learning_rate": 8.835793649683342e-05,
+      "loss": 3.5572,
+      "step": 3800
+    },
+    {
+      "epoch": 0.11732719621845421,
+      "grad_norm": 0.30023226141929626,
+      "learning_rate": 8.770809879244541e-05,
+      "loss": 3.5447,
+      "step": 3900
+    },
+    {
+      "epoch": 0.12033558586508125,
+      "grad_norm": 0.2889101207256317,
+      "learning_rate": 8.704314379665195e-05,
+      "loss": 3.5334,
+      "step": 4000
+    },
+    {
+      "epoch": 0.12322363992584319,
+      "eval_loss": 3.5199642181396484,
+      "eval_runtime": 25.5123,
+      "eval_samples_per_second": 166.116,
+      "eval_steps_per_second": 5.213,
+      "step": 4096
+    },
+    {
+      "epoch": 0.12334397551170828,
+      "grad_norm": 0.2787327468395233,
+      "learning_rate": 8.636333809193176e-05,
+      "loss": 3.5224,
+      "step": 4100
+    },
+    {
+      "epoch": 0.1263523651583353,
+      "grad_norm": 0.28778326511383057,
+      "learning_rate": 8.566895421445784e-05,
+      "loss": 3.5129,
+      "step": 4200
+    },
+    {
+      "epoch": 0.12936075480496234,
+      "grad_norm": 0.2771563231945038,
+      "learning_rate": 8.496027054483702e-05,
+      "loss": 3.5002,
+      "step": 4300
+    },
+    {
+      "epoch": 0.13236914445158937,
+      "grad_norm": 0.28959691524505615,
+      "learning_rate": 8.423757119650619e-05,
+      "loss": 3.4917,
+      "step": 4400
+    },
+    {
+      "epoch": 0.1353775340982164,
+      "grad_norm": 0.27935898303985596,
+      "learning_rate": 8.35011459018305e-05,
+      "loss": 3.4875,
+      "step": 4500
+    },
+    {
+      "epoch": 0.13838592374484343,
+      "grad_norm": 0.2650796175003052,
+      "learning_rate": 8.275128989594867e-05,
+      "loss": 3.4685,
+      "step": 4600
+    },
+    {
+      "epoch": 0.14139431339147046,
+      "grad_norm": 0.2627846598625183,
+      "learning_rate": 8.198830379841233e-05,
+      "loss": 3.4681,
+      "step": 4700
+    },
+    {
+      "epoch": 0.1444027030380975,
+      "grad_norm": 0.28169551491737366,
+      "learning_rate": 8.121249349266684e-05,
+      "loss": 3.4556,
+      "step": 4800
+    },
+    {
+      "epoch": 0.14741109268472452,
+      "grad_norm": 0.2804490327835083,
+      "learning_rate": 8.042417000342151e-05,
+      "loss": 3.4503,
+      "step": 4900
+    },
+    {
+      "epoch": 0.15041948233135155,
+      "grad_norm": 0.3104579746723175,
+      "learning_rate": 7.962364937195902e-05,
+      "loss": 3.4394,
+      "step": 5000
+    },
+    {
+      "epoch": 0.15342787197797858,
+      "grad_norm": 0.27914077043533325,
+      "learning_rate": 7.88112525294334e-05,
+      "loss": 3.4354,
+      "step": 5100
+    },
+    {
+      "epoch": 0.154029549907304,
+      "eval_loss": 3.428358554840088,
+      "eval_runtime": 25.6005,
+      "eval_samples_per_second": 165.544,
+      "eval_steps_per_second": 5.195,
+      "step": 5120
+    },
+    {
+      "epoch": 0.1564362616246056,
+      "grad_norm": 0.2609664499759674,
+      "learning_rate": 7.798730516820779e-05,
+      "loss": 3.4221,
+      "step": 5200
+    },
+    {
+      "epoch": 0.15944465127123264,
+      "grad_norm": 0.26283183693885803,
+      "learning_rate": 7.715213761128337e-05,
+      "loss": 3.4155,
+      "step": 5300
+    },
+    {
+      "epoch": 0.16245304091785967,
+      "grad_norm": 0.27422934770584106,
+      "learning_rate": 7.63060846798719e-05,
+      "loss": 3.4033,
+      "step": 5400
+    },
+    {
+      "epoch": 0.1654614305644867,
+      "grad_norm": 0.28960829973220825,
+      "learning_rate": 7.544948555916485e-05,
+      "loss": 3.4011,
+      "step": 5500
+    },
+    {
+      "epoch": 0.16846982021111373,
+      "grad_norm": 0.26780277490615845,
+      "learning_rate": 7.4582683662353e-05,
+      "loss": 3.3995,
+      "step": 5600
+    },
+    {
+      "epoch": 0.1714782098577408,
+      "grad_norm": 0.26375481486320496,
+      "learning_rate": 7.370602649295112e-05,
+      "loss": 3.3901,
+      "step": 5700
+    },
+    {
+      "epoch": 0.17448659950436782,
+      "grad_norm": 0.26001670956611633,
+      "learning_rate": 7.281986550548268e-05,
+      "loss": 3.3838,
+      "step": 5800
+    },
+    {
+      "epoch": 0.17749498915099485,
+      "grad_norm": 0.2617439925670624,
+      "learning_rate": 7.192455596458073e-05,
+      "loss": 3.3786,
+      "step": 5900
+    },
+    {
+      "epoch": 0.18050337879762188,
+      "grad_norm": 0.2578573226928711,
+      "learning_rate": 7.102045680256113e-05,
+      "loss": 3.3686,
+      "step": 6000
+    },
+    {
+      "epoch": 0.1835117684442489,
+      "grad_norm": 0.25970160961151123,
+      "learning_rate": 7.010793047552565e-05,
+      "loss": 3.3599,
+      "step": 6100
+    },
+    {
+      "epoch": 0.1848354598887648,
+      "eval_loss": 3.3594183921813965,
+      "eval_runtime": 28.2705,
+      "eval_samples_per_second": 149.909,
+      "eval_steps_per_second": 4.705,
+      "step": 6144
+    },
+    {
+      "epoch": 0.18652015809087594,
+      "grad_norm": 0.25814658403396606,
+      "learning_rate": 6.918734281805204e-05,
+      "loss": 3.3542,
+      "step": 6200
+    },
+    {
+      "epoch": 0.18952854773750297,
+      "grad_norm": 0.26643288135528564,
+      "learning_rate": 6.825906289652998e-05,
+      "loss": 3.3486,
+      "step": 6300
+    },
+    {
+      "epoch": 0.19253693738413,
+      "grad_norm": 0.2631397545337677,
+      "learning_rate": 6.732346286120116e-05,
+      "loss": 3.3471,
+      "step": 6400
+    },
+    {
+      "epoch": 0.19554532703075703,
+      "grad_norm": 0.25272083282470703,
+      "learning_rate": 6.638091779696314e-05,
+      "loss": 3.3379,
+      "step": 6500
+    },
+    {
+      "epoch": 0.19855371667738406,
+      "grad_norm": 0.2605939507484436,
+      "learning_rate": 6.543180557299673e-05,
+      "loss": 3.3298,
+      "step": 6600
+    },
+    {
+      "epoch": 0.2015621063240111,
+      "grad_norm": 0.25710055232048035,
+      "learning_rate": 6.447650669127692e-05,
+      "loss": 3.3237,
+      "step": 6700
+    },
+    {
+      "epoch": 0.20457049597063812,
+      "grad_norm": 0.2481650710105896,
+      "learning_rate": 6.35154041340287e-05,
+      "loss": 3.3253,
+      "step": 6800
+    },
+    {
+      "epoch": 0.20757888561726515,
+      "grad_norm": 0.37661871314048767,
+      "learning_rate": 6.254888321018817e-05,
+      "loss": 3.3143,
+      "step": 6900
+    },
+    {
+      "epoch": 0.21058727526389218,
+      "grad_norm": 0.2581317722797394,
+      "learning_rate": 6.157733140093107e-05,
+      "loss": 3.3149,
+      "step": 7000
+    },
+    {
+      "epoch": 0.2135956649105192,
+      "grad_norm": 0.25675368309020996,
+      "learning_rate": 6.060113820433041e-05,
+      "loss": 3.3015,
+      "step": 7100
+    },
+    {
+      "epoch": 0.21564136987022559,
+      "eval_loss": 3.304588794708252,
+      "eval_runtime": 27.5945,
+      "eval_samples_per_second": 153.581,
+      "eval_steps_per_second": 4.82,
+      "step": 7168
+    },
+    {
+      "epoch": 0.21660405455714624,
+      "grad_norm": 0.25401395559310913,
+      "learning_rate": 5.962069497920546e-05,
+      "loss": 3.302,
+      "step": 7200
+    },
+    {
+      "epoch": 0.21961244420377327,
+      "grad_norm": 0.2600981295108795,
+      "learning_rate": 5.863639478822488e-05,
+      "loss": 3.2984,
+      "step": 7300
+    },
+    {
+      "epoch": 0.2226208338504003,
+      "grad_norm": 0.253537654876709,
+      "learning_rate": 5.764863224032663e-05,
+      "loss": 3.2987,
+      "step": 7400
+    },
+    {
+      "epoch": 0.22562922349702733,
+      "grad_norm": 0.2553566098213196,
+      "learning_rate": 5.66578033325181e-05,
+      "loss": 3.2817,
+      "step": 7500
+    },
+    {
+      "epoch": 0.22863761314365436,
+      "grad_norm": 0.2590862810611725,
+      "learning_rate": 5.566430529111969e-05,
+      "loss": 3.2848,
+      "step": 7600
+    },
+    {
+      "epoch": 0.23164600279028139,
+      "grad_norm": 0.24905967712402344,
+      "learning_rate": 5.4668536412515534e-05,
+      "loss": 3.2758,
+      "step": 7700
+    },
+    {
+      "epoch": 0.23465439243690842,
+      "grad_norm": 0.25471484661102295,
+      "learning_rate": 5.367089590347535e-05,
+      "loss": 3.2805,
+      "step": 7800
+    },
+    {
+      "epoch": 0.23766278208353547,
+      "grad_norm": 0.25363436341285706,
+      "learning_rate": 5.2671783721111145e-05,
+      "loss": 3.274,
+      "step": 7900
+    },
+    {
+      "epoch": 0.2406711717301625,
+      "grad_norm": 0.2578498423099518,
+      "learning_rate": 5.1671600412533284e-05,
+      "loss": 3.2648,
+      "step": 8000
+    },
+    {
+      "epoch": 0.24367956137678953,
+      "grad_norm": 0.2545549273490906,
+      "learning_rate": 5.067074695426984e-05,
+      "loss": 3.259,
+      "step": 8100
+    },
+    {
+      "epoch": 0.24644727985168638,
+      "eval_loss": 3.2613685131073,
+      "eval_runtime": 28.3851,
+      "eval_samples_per_second": 149.304,
+      "eval_steps_per_second": 4.686,
+      "step": 8192
+    }
+  ],
+  "logging_steps": 100,
+  "max_steps": 16011,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 2048,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.0944796278785049e+19,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0301c2ef51a631de0b3b6ebc6c3d958b7f37efc6d6c6e0bd2902aed002e99e04
+size 7544

checkpoints/qwen3/qwen3_p1083m_t1_2_suite-owt_scaling_v3_family-qwen3_geometry_id-euclid_variant-base_init_slope-std/attempt2_20260219_093238/checkpoint-8192/zero_to_fp32.py

@@ -0,0 +1,604 @@
+#!/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 . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+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)
+
+        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 files:
+        state_dict = torch.load(f, map_location=device)
+        # 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 not ZERO_STAGE 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}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) 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")
+
+
+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 = fp32_flat_groups[0].numel() * 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
+    for name, shape in param_shapes.items():
+
+        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}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        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 get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=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
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    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.
+
+    """
+    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")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, 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_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+        - ``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
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+    print(f"Saving fp32 state dict to {output_file}")
+    torch.save(state_dict, output_file)
+
+
+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(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"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_file",
+        type=str,
+        help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+    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_file,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)