EPFL final checkpoints and W&B history 351-400 of 1463

.gitattributes

@@ -43,3 +43,10 @@ epfl_final/SparseAtt/wandb/run-20260707_035814-kxni9nyw/run-kxni9nyw.wandb filte
 epfl_final/SparseAtt/wandb/run-20260707_035818-m10lf13v/run-m10lf13v.wandb filter=lfs diff=lfs merge=lfs -text
 epfl_final/SparseAtt/wandb/run-20260707_035828-jmi9928p/run-jmi9928p.wandb filter=lfs diff=lfs merge=lfs -text
 epfl_final/SparseAtt/wandb/run-20260707_035835-sh1m3lmh/run-sh1m3lmh.wandb filter=lfs diff=lfs merge=lfs -text
+epfl_final/SparseAtt/wandb/run-20260707_170941-qlqhc3gy/run-qlqhc3gy.wandb filter=lfs diff=lfs merge=lfs -text
+epfl_final/SparseAtt/wandb/run-20260707_170941-dbtl6d2o/run-dbtl6d2o.wandb filter=lfs diff=lfs merge=lfs -text
+epfl_final/SparseAtt/wandb/run-20260707_170941-lgjoqt2r/run-lgjoqt2r.wandb filter=lfs diff=lfs merge=lfs -text
+epfl_final/SparseAtt/wandb/run-20260707_171131-304r0ilz/run-304r0ilz.wandb filter=lfs diff=lfs merge=lfs -text
+epfl_final/SparseAtt/wandb/run-20260708_045920-qlqhc3gy/run-qlqhc3gy.wandb filter=lfs diff=lfs merge=lfs -text
+epfl_final/SparseAtt/wandb/run-20260708_054405-43sa4ouk/run-43sa4ouk.wandb filter=lfs diff=lfs merge=lfs -text
+epfl_final/SparseAtt/wandb/run-20260708_060003-qlqhc3gy/run-qlqhc3gy.wandb filter=lfs diff=lfs merge=lfs -text

epfl_final/SparseAtt/wandb/run-20260707_170941-dbtl6d2o/files/config.yaml

@@ -0,0 +1,189 @@
+_wandb:
+    value:
+        cli_version: 0.26.1
+        e:
+            p578vnxv2tp5uy8pd7xp1mgwafzl15g5:
+                args:
+                    - --config
+                    - configs/ladder8k/fixed_s1.json
+                codePath: train.py
+                codePathLocal: train.py
+                cpu_count: 96
+                cpu_count_logical: 192
+                cudaVersion: "13.0"
+                disk:
+                    /:
+                        total: "9371709865984"
+                        used: "5840433963008"
+                email: ruben.navasardyan@mail.utoronto.ca
+                executable: /mloscratch/homes/navasard/conda/bin/python
+                git:
+                    commit: 88298c7a418622530ca5426b22b1e5ebd10ca0cb
+                    remote: https://github.com/rubentium/TokenPrune.git
+                gpu: NVIDIA H100 80GB HBM3
+                gpu_count: 4
+                gpu_nvidia:
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-98029aa4-c589-b65b-3ab7-6495256f31ed
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-37a1c6c2-d3df-3064-44df-33fa82304bd0
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-b87e8111-63b6-b364-28b9-c4ce747aec08
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-bbbaaf81-9234-a1c6-0093-6d87b4e6af4c
+                host: h100-workspace-4-gpu-0-4
+                memory:
+                    total: "1623005732864"
+                os: Linux-6.8.0-124-generic-x86_64-with-glibc2.35
+                program: /mloscratch/homes/navasard/sparse/SparseAtt/train.py
+                python: CPython 3.13.13
+                root: /mloscratch/homes/navasard/sparse/SparseAtt
+                startedAt: "2026-07-07T17:09:41.400774Z"
+                writerId: p578vnxv2tp5uy8pd7xp1mgwafzl15g5
+        m: []
+        python_version: 3.13.13
+        t:
+            "1":
+                - 1
+                - 5
+                - 11
+                - 49
+                - 51
+                - 53
+                - 71
+            "2":
+                - 1
+                - 5
+                - 11
+                - 49
+                - 51
+                - 53
+                - 71
+            "3":
+                - 2
+                - 5
+                - 13
+                - 14
+                - 15
+                - 16
+                - 61
+                - 62
+            "4": 3.13.13
+            "5": 0.26.1
+            "6": 5.7.0
+            "12": 0.26.1
+            "13": linux-x86_64
+attention_type:
+    value: sparse_stochastic
+checkpoint_activations:
+    value: false
+dense_mode:
+    value: false
+dense_step_intervals:
+    value: null
+eval_batches:
+    value: 50
+eval_steps:
+    value: 250
+force_new_run:
+    value: false
+gradient_accumulation_steps:
+    value: 10
+head_dim:
+    value: null
+hidden_size:
+    value: 768
+init_from_scratch:
+    value: true
+intermediate_size:
+    value: 3072
+learning_rate:
+    value: 0.0005
+logging_steps:
+    value: 10
+lr_scheduler:
+    value: cosine
+max_grad_norm:
+    value: 1
+max_seq_length:
+    value: 8192
+max_steps:
+    value: 44000
+mixed_precision:
+    value: bf16
+model_dtype:
+    value: bfloat16
+model_path:
+    value: /mloscratch/homes/navasard/qwen3_06b_base
+num_attention_heads:
+    value: 6
+num_epochs:
+    value: 1
+num_hidden_layers:
+    value: 28
+num_key_value_heads:
+    value: 2
+num_workers:
+    value: 8
+output_dir:
+    value: /mloscratch/homes/navasard/sparse/checkpoints/l8k365m_fixed_s1/run_20260707_034557
+per_device_batch_size:
+    value: 2
+resume_from_checkpoint:
+    value: /mloscratch/homes/navasard/sparse/checkpoints/l8k365m_fixed_s1/run_20260707_034557/checkpoint-14000
+save_steps:
+    value: 1000
+seed:
+    value: 42
+sparse_block_size:
+    value: 128
+sparse_layer_indices:
+    value: null
+sparse_mask_cache_dir:
+    value: /mloscratch/homes/navasard/sparse/mask_cache_8k_fixed_s1
+sparse_mask_pool_size:
+    value: 1
+sparse_mask_seed:
+    value: 1
+sparse_num_random_blocks:
+    value: 31
+sparse_val_dense_compare:
+    value: false
+strategy:
+    value: ddp
+train_path:
+    value: /mloscratch/homes/navasard/data/fineweb_large/tokenized
+val_path:
+    value: null
+verbose:
+    value: true
+wandb_enabled:
+    value: true
+wandb_name:
+    value: 365m fixed s1 8k
+wandb_project:
+    value: stochastic_sparsity
+wandb_tags:
+    value:
+        - qwen3
+        - 365m
+        - 8k
+        - scaleladder
+        - fixed
+        - seed1
+warmup_steps:
+    value: 750
+weight_decay:
+    value: 0.01

epfl_final/SparseAtt/wandb/run-20260707_170941-dbtl6d2o/files/requirements.txt

@@ -0,0 +1,141 @@
+pip==26.0.1
+nvidia-cusparselt-cu13==0.8.0
+mpmath==1.3.0
+cuda-toolkit==13.0.2
+zstandard==0.25.0
+xxhash==3.7.0
+urllib3==2.6.3
+typing_extensions==4.15.0
+triton==3.6.0
+tqdm==4.67.3
+torchinfo==1.8.0
+sympy==1.14.0
+smmap==5.0.3
+six==1.17.0
+shellingham==1.5.4
+setuptools==81.0.0
+safetensors==0.7.0
+regex==2026.4.4
+PyYAML==6.0.3
+Pygments==2.20.0
+pyarrow==24.0.0
+psutil==7.2.2
+protobuf==7.34.1
+propcache==0.4.1
+platformdirs==4.9.6
+packaging==26.2
+nvidia-nvtx==13.0.85
+nvidia-nvshmem-cu13==3.4.5
+nvidia-nvjitlink==13.0.88
+nvidia-nccl-cu13==2.28.9
+nvidia-curand==10.4.0.35
+nvidia-cufile==1.15.1.6
+nvidia-cuda-runtime==13.0.96
+nvidia-cuda-nvrtc==13.0.88
+nvidia-cuda-cupti==13.0.85
+nvidia-cublas==13.1.0.3
+numpy==2.4.4
+networkx==3.6.1
+multidict==6.7.1
+mdurl==0.1.2
+MarkupSafe==3.0.3
+idna==3.13
+hf-xet==1.4.3
+h11==0.16.0
+fsspec==2026.2.0
+frozenlist==1.8.0
+filelock==3.29.0
+dill==0.4.1
+cuda-pathfinder==1.5.4
+click==8.3.3
+charset-normalizer==3.4.7
+certifi==2026.4.22
+attrs==26.1.0
+annotated-types==0.7.0
+annotated-doc==0.0.4
+aiohappyeyeballs==2.6.1
+yarl==1.23.0
+typing-inspection==0.4.2
+sentry-sdk==2.58.0
+requests==2.33.1
+python-dateutil==2.9.0.post0
+pydantic_core==2.46.3
+nvidia-cusparse==12.6.3.3
+nvidia-cufft==12.0.0.61
+nvidia-cudnn-cu13==9.19.0.56
+multiprocess==0.70.19
+markdown-it-py==4.0.0
+Jinja2==3.1.6
+httpcore==1.0.9
+gitdb==4.0.12
+cuda-bindings==13.2.0
+anyio==4.13.0
+aiosignal==1.4.0
+rich==15.0.0
+pydantic==2.13.3
+pandas==3.0.2
+nvidia-cusolver==12.0.4.66
+httpx==0.28.1
+GitPython==3.1.49
+aiohttp==3.13.5
+wandb==0.26.1
+typer==0.25.1
+torch==2.11.0
+huggingface_hub==1.13.0
+tokenizers==0.22.2
+datasets==4.8.5
+accelerate==1.13.0
+transformers==5.7.0
+word2number==1.1
+sqlitedict==2.1.0
+pytz==2026.2
+threadpoolctl==3.6.0
+tcolorpy==0.1.7
+tabulate==0.10.0
+scipy==1.17.1
+portalocker==3.2.0
+pathvalidate==3.3.1
+narwhals==2.22.1
+more-itertools==11.1.0
+lxml==6.1.1
+joblib==1.5.3
+colorama==0.4.6
+chardet==6.0.0.post1
+absl-py==2.4.0
+scikit-learn==1.9.0
+sacrebleu==2.6.0
+nltk==3.9.4
+mbstrdecoder==1.1.5
+typepy==1.3.5
+rouge_score==0.1.2
+DataProperty==1.1.1
+tabledata==1.3.5
+evaluate==0.4.6
+pytablewriter==1.2.1
+lm_eval==0.4.12
+pluggy==1.6.0
+iniconfig==2.3.0
+pytest==9.1.0
+py-spy==0.4.2
+pillow==12.2.0
+tifffile==2026.6.1
+imagecodecs==2026.6.26
+zipp==4.1.0
+importlib_metadata==9.0.0
+derivative==0.6.3
+pysindy==2.1.0
+pyparsing==3.3.2
+kiwisolver==1.5.0
+fonttools==4.63.0
+cycler==0.12.1
+contourpy==1.3.3
+matplotlib==3.11.0
+sortedcontainers==2.4.0
+wcwidth==0.8.2
+sigfig==1.3.19
+importlib_resources==7.1.0
+enum-tools==0.13.0
+cramjam==2.11.0
+python-snappy==0.7.3
+compact-json==1.8.2
+numbers-parser==4.18.5

epfl_final/SparseAtt/wandb/run-20260707_170941-dbtl6d2o/files/wandb-metadata.json

@@ -0,0 +1,65 @@
+{
+  "os":  "Linux-6.8.0-124-generic-x86_64-with-glibc2.35",
+  "python":  "CPython 3.13.13",
+  "startedAt":  "2026-07-07T17:09:41.400774Z",
+  "args":  [
+    "--config",
+    "configs/ladder8k/fixed_s1.json"
+  ],
+  "program":  "/mloscratch/homes/navasard/sparse/SparseAtt/train.py",
+  "codePath":  "train.py",
+  "codePathLocal":  "train.py",
+  "git":  {
+    "remote":  "https://github.com/rubentium/TokenPrune.git",
+    "commit":  "88298c7a418622530ca5426b22b1e5ebd10ca0cb"
+  },
+  "email":  "ruben.navasardyan@mail.utoronto.ca",
+  "root":  "/mloscratch/homes/navasard/sparse/SparseAtt",
+  "host":  "h100-workspace-4-gpu-0-4",
+  "executable":  "/mloscratch/homes/navasard/conda/bin/python",
+  "cpu_count":  96,
+  "cpu_count_logical":  192,
+  "gpu":  "NVIDIA H100 80GB HBM3",
+  "gpu_count":  4,
+  "disk":  {
+    "/":  {
+      "total":  "9371709865984",
+      "used":  "5840433963008"
+    }
+  },
+  "memory":  {
+    "total":  "1623005732864"
+  },
+  "gpu_nvidia":  [
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-98029aa4-c589-b65b-3ab7-6495256f31ed"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-37a1c6c2-d3df-3064-44df-33fa82304bd0"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-b87e8111-63b6-b364-28b9-c4ce747aec08"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-bbbaaf81-9234-a1c6-0093-6d87b4e6af4c"
+    }
+  ],
+  "cudaVersion":  "13.0",
+  "writerId":  "p578vnxv2tp5uy8pd7xp1mgwafzl15g5"
+}

epfl_final/SparseAtt/wandb/run-20260707_170941-dbtl6d2o/files/wandb-summary.json

@@ -0,0 +1 @@
+{"_timestamp":1.783485718332791e+09,"_step":44000,"speed/steps_per_sec":0.7784842981037987,"eval/dense_interval_active":0,"train/global_step":44000,"eval/uses_sparse_attention":1,"_runtime":61965.56091351,"eval/dense_interval_end":-1,"train/loss":2.9672227907180786,"speed/ms_per_step":1284.5474243164062,"_wandb":{"runtime":61965},"train/epoch":0.2906439083679023,"eval/perplexity":18.80757783227063,"eval/dense_interval_start":-1,"speed/samples_per_sec":15.569685962075974,"eval/loss":2.934259864807129,"eval/uses_dense_attention":0,"train/learning_rate":0}

epfl_final/SparseAtt/wandb/run-20260707_170941-dbtl6d2o/logs/debug.log

@@ -0,0 +1,25 @@
+2026-07-07 17:09:41,432 INFO    MainThread:431 [wandb_setup.py:_flush():81] Current SDK version is 0.26.1
+2026-07-07 17:09:41,433 INFO    MainThread:431 [wandb_setup.py:_flush():81] Configure stats pid to 431
+2026-07-07 17:09:41,433 INFO    MainThread:431 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2026-07-07 17:09:41,433 INFO    MainThread:431 [wandb_init.py:setup_run_log_directory():723] Logging user logs to /mloscratch/homes/navasard/sparse/SparseAtt/wandb/run-20260707_170941-dbtl6d2o/logs/debug.log
+2026-07-07 17:09:41,433 INFO    MainThread:431 [wandb_init.py:setup_run_log_directory():724] Logging internal logs to /mloscratch/homes/navasard/sparse/SparseAtt/wandb/run-20260707_170941-dbtl6d2o/logs/debug-internal.log
+2026-07-07 17:09:41,433 INFO    MainThread:431 [wandb_init.py:init():850] calling init triggers
+2026-07-07 17:09:41,433 INFO    MainThread:431 [wandb_init.py:init():855] wandb.init called with sweep_config: {}
+config: {'model_path': '/mloscratch/homes/navasard/qwen3_06b_base', 'model_dtype': 'bfloat16', 'init_from_scratch': True, 'train_path': '/mloscratch/homes/navasard/data/fineweb_large/tokenized', 'val_path': None, 'max_seq_length': 8192, 'num_workers': 8, 'output_dir': '/mloscratch/homes/navasard/sparse/checkpoints/l8k365m_fixed_s1/run_20260707_034557', 'resume_from_checkpoint': '/mloscratch/homes/navasard/sparse/checkpoints/l8k365m_fixed_s1/run_20260707_034557/checkpoint-14000', 'force_new_run': False, 'num_epochs': 1, 'max_steps': 44000, 'per_device_batch_size': 2, 'gradient_accumulation_steps': 10, 'learning_rate': 0.0005, 'weight_decay': 0.01, 'warmup_steps': 750, 'lr_scheduler': 'cosine', 'max_grad_norm': 1.0, 'save_steps': 1000, 'eval_steps': 250, 'eval_batches': 50, 'logging_steps': 10, 'seed': 42, 'verbose': True, 'checkpoint_activations': False, 'strategy': 'ddp', 'mixed_precision': 'bf16', 'wandb_enabled': True, 'wandb_project': 'stochastic_sparsity', 'wandb_name': '365m fixed s1 8k', 'wandb_tags': ['qwen3', '365m', '8k', 'scaleladder', 'fixed', 'seed1'], 'hidden_size': 768, 'intermediate_size': 3072, 'num_attention_heads': 6, 'num_key_value_heads': 2, 'num_hidden_layers': 28, 'head_dim': None, 'attention_type': 'sparse_stochastic', 'sparse_block_size': 128, 'sparse_num_random_blocks': 31, 'sparse_mask_pool_size': 1, 'sparse_val_dense_compare': False, 'sparse_layer_indices': None, 'sparse_mask_cache_dir': '/mloscratch/homes/navasard/sparse/mask_cache_8k_fixed_s1', 'sparse_mask_seed': 1, 'dense_mode': False, 'dense_step_intervals': None, '_wandb': {}}
+2026-07-07 17:09:41,433 INFO    MainThread:431 [wandb_init.py:init():898] starting backend
+2026-07-07 17:09:41,986 INFO    MainThread:431 [wandb_init.py:init():913] sending inform_init request
+2026-07-07 17:09:42,434 INFO    MainThread:431 [wandb_init.py:init():918] backend started and connected
+2026-07-07 17:09:42,439 INFO    MainThread:431 [wandb_init.py:init():988] updated telemetry
+2026-07-07 17:09:42,460 INFO    MainThread:431 [wandb_init.py:init():1011] communicating run to backend with 90.0 second timeout
+2026-07-07 17:09:43,071 INFO    MainThread:431 [wandb_init.py:init():1051] run resumed
+2026-07-07 17:09:43,073 INFO    MainThread:431 [wandb_init.py:init():1056] starting run threads in backend
+2026-07-07 17:09:43,318 INFO    MainThread:431 [wandb_run.py:_console_start():2554] atexit reg
+2026-07-07 17:09:43,318 INFO    MainThread:431 [wandb_run.py:_redirect():2403] redirect: wrap_raw
+2026-07-07 17:09:43,318 INFO    MainThread:431 [wandb_run.py:_redirect():2472] Wrapping output streams.
+2026-07-07 17:09:43,318 INFO    MainThread:431 [wandb_run.py:_redirect():2495] Redirects installed.
+2026-07-07 17:09:43,322 INFO    MainThread:431 [wandb_init.py:init():1094] run started, returning control to user process
+2026-07-08 04:42:04,632 INFO    MainThread:431 [wandb_run.py:_finish():2321] finishing run ruben-navasardyan4869-epfl/stochastic_sparsity/dbtl6d2o
+2026-07-08 04:42:04,633 INFO    MainThread:431 [wandb_run.py:_atexit_cleanup():2520] got exitcode: 0
+2026-07-08 04:42:04,633 INFO    MainThread:431 [wandb_run.py:_restore():2502] restore
+2026-07-08 04:42:04,633 INFO    MainThread:431 [wandb_run.py:_restore():2508] restore done
+2026-07-08 04:42:06,596 INFO    MainThread:431 [wandb_run.py:_footer_sync_info():3889] logging synced files

epfl_final/SparseAtt/wandb/run-20260707_170941-dbtl6d2o/run-dbtl6d2o.wandb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cbc9e3d922cd26f4d7dd7e760d2e20bc1cce156d858ea0fc0a62e18e5795521c
+size 15386389

epfl_final/SparseAtt/wandb/run-20260707_170941-lgjoqt2r/files/config.yaml

@@ -0,0 +1,189 @@
+_wandb:
+    value:
+        cli_version: 0.26.1
+        e:
+            1poap8apibhpfjc9hz6ylrlyso44bwkc:
+                args:
+                    - --config
+                    - configs/ladder8k/fixed_s2.json
+                codePath: train.py
+                codePathLocal: train.py
+                cpu_count: 96
+                cpu_count_logical: 192
+                cudaVersion: "13.0"
+                disk:
+                    /:
+                        total: "9371709865984"
+                        used: "5840433963008"
+                email: ruben.navasardyan@mail.utoronto.ca
+                executable: /mloscratch/homes/navasard/conda/bin/python
+                git:
+                    commit: 88298c7a418622530ca5426b22b1e5ebd10ca0cb
+                    remote: https://github.com/rubentium/TokenPrune.git
+                gpu: NVIDIA H100 80GB HBM3
+                gpu_count: 4
+                gpu_nvidia:
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-98029aa4-c589-b65b-3ab7-6495256f31ed
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-37a1c6c2-d3df-3064-44df-33fa82304bd0
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-b87e8111-63b6-b364-28b9-c4ce747aec08
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-bbbaaf81-9234-a1c6-0093-6d87b4e6af4c
+                host: h100-workspace-4-gpu-0-4
+                memory:
+                    total: "1623005732864"
+                os: Linux-6.8.0-124-generic-x86_64-with-glibc2.35
+                program: /mloscratch/homes/navasard/sparse/SparseAtt/train.py
+                python: CPython 3.13.13
+                root: /mloscratch/homes/navasard/sparse/SparseAtt
+                startedAt: "2026-07-07T17:09:41.403720Z"
+                writerId: 1poap8apibhpfjc9hz6ylrlyso44bwkc
+        m: []
+        python_version: 3.13.13
+        t:
+            "1":
+                - 1
+                - 5
+                - 11
+                - 49
+                - 51
+                - 53
+                - 71
+            "2":
+                - 1
+                - 5
+                - 11
+                - 49
+                - 51
+                - 53
+                - 71
+            "3":
+                - 2
+                - 5
+                - 13
+                - 14
+                - 15
+                - 16
+                - 61
+                - 62
+            "4": 3.13.13
+            "5": 0.26.1
+            "6": 5.7.0
+            "12": 0.26.1
+            "13": linux-x86_64
+attention_type:
+    value: sparse_stochastic
+checkpoint_activations:
+    value: false
+dense_mode:
+    value: false
+dense_step_intervals:
+    value: null
+eval_batches:
+    value: 50
+eval_steps:
+    value: 250
+force_new_run:
+    value: false
+gradient_accumulation_steps:
+    value: 10
+head_dim:
+    value: null
+hidden_size:
+    value: 768
+init_from_scratch:
+    value: true
+intermediate_size:
+    value: 3072
+learning_rate:
+    value: 0.0005
+logging_steps:
+    value: 10
+lr_scheduler:
+    value: cosine
+max_grad_norm:
+    value: 1
+max_seq_length:
+    value: 8192
+max_steps:
+    value: 44000
+mixed_precision:
+    value: bf16
+model_dtype:
+    value: bfloat16
+model_path:
+    value: /mloscratch/homes/navasard/qwen3_06b_base
+num_attention_heads:
+    value: 6
+num_epochs:
+    value: 1
+num_hidden_layers:
+    value: 28
+num_key_value_heads:
+    value: 2
+num_workers:
+    value: 8
+output_dir:
+    value: /mloscratch/homes/navasard/sparse/checkpoints/l8k365m_fixed_s2/run_20260707_034600
+per_device_batch_size:
+    value: 2
+resume_from_checkpoint:
+    value: /mloscratch/homes/navasard/sparse/checkpoints/l8k365m_fixed_s2/run_20260707_034600/checkpoint-15000
+save_steps:
+    value: 1000
+seed:
+    value: 42
+sparse_block_size:
+    value: 128
+sparse_layer_indices:
+    value: null
+sparse_mask_cache_dir:
+    value: /mloscratch/homes/navasard/sparse/mask_cache_8k_fixed_s2
+sparse_mask_pool_size:
+    value: 1
+sparse_mask_seed:
+    value: 2
+sparse_num_random_blocks:
+    value: 31
+sparse_val_dense_compare:
+    value: false
+strategy:
+    value: ddp
+train_path:
+    value: /mloscratch/homes/navasard/data/fineweb_large/tokenized
+val_path:
+    value: null
+verbose:
+    value: true
+wandb_enabled:
+    value: true
+wandb_name:
+    value: 365m fixed s2 8k
+wandb_project:
+    value: stochastic_sparsity
+wandb_tags:
+    value:
+        - qwen3
+        - 365m
+        - 8k
+        - scaleladder
+        - fixed
+        - seed2
+warmup_steps:
+    value: 750
+weight_decay:
+    value: 0.01

epfl_final/SparseAtt/wandb/run-20260707_170941-lgjoqt2r/files/requirements.txt

@@ -0,0 +1,141 @@
+pip==26.0.1
+nvidia-cusparselt-cu13==0.8.0
+mpmath==1.3.0
+cuda-toolkit==13.0.2
+zstandard==0.25.0
+xxhash==3.7.0
+urllib3==2.6.3
+typing_extensions==4.15.0
+triton==3.6.0
+tqdm==4.67.3
+torchinfo==1.8.0
+sympy==1.14.0
+smmap==5.0.3
+six==1.17.0
+shellingham==1.5.4
+setuptools==81.0.0
+safetensors==0.7.0
+regex==2026.4.4
+PyYAML==6.0.3
+Pygments==2.20.0
+pyarrow==24.0.0
+psutil==7.2.2
+protobuf==7.34.1
+propcache==0.4.1
+platformdirs==4.9.6
+packaging==26.2
+nvidia-nvtx==13.0.85
+nvidia-nvshmem-cu13==3.4.5
+nvidia-nvjitlink==13.0.88
+nvidia-nccl-cu13==2.28.9
+nvidia-curand==10.4.0.35
+nvidia-cufile==1.15.1.6
+nvidia-cuda-runtime==13.0.96
+nvidia-cuda-nvrtc==13.0.88
+nvidia-cuda-cupti==13.0.85
+nvidia-cublas==13.1.0.3
+numpy==2.4.4
+networkx==3.6.1
+multidict==6.7.1
+mdurl==0.1.2
+MarkupSafe==3.0.3
+idna==3.13
+hf-xet==1.4.3
+h11==0.16.0
+fsspec==2026.2.0
+frozenlist==1.8.0
+filelock==3.29.0
+dill==0.4.1
+cuda-pathfinder==1.5.4
+click==8.3.3
+charset-normalizer==3.4.7
+certifi==2026.4.22
+attrs==26.1.0
+annotated-types==0.7.0
+annotated-doc==0.0.4
+aiohappyeyeballs==2.6.1
+yarl==1.23.0
+typing-inspection==0.4.2
+sentry-sdk==2.58.0
+requests==2.33.1
+python-dateutil==2.9.0.post0
+pydantic_core==2.46.3
+nvidia-cusparse==12.6.3.3
+nvidia-cufft==12.0.0.61
+nvidia-cudnn-cu13==9.19.0.56
+multiprocess==0.70.19
+markdown-it-py==4.0.0
+Jinja2==3.1.6
+httpcore==1.0.9
+gitdb==4.0.12
+cuda-bindings==13.2.0
+anyio==4.13.0
+aiosignal==1.4.0
+rich==15.0.0
+pydantic==2.13.3
+pandas==3.0.2
+nvidia-cusolver==12.0.4.66
+httpx==0.28.1
+GitPython==3.1.49
+aiohttp==3.13.5
+wandb==0.26.1
+typer==0.25.1
+torch==2.11.0
+huggingface_hub==1.13.0
+tokenizers==0.22.2
+datasets==4.8.5
+accelerate==1.13.0
+transformers==5.7.0
+word2number==1.1
+sqlitedict==2.1.0
+pytz==2026.2
+threadpoolctl==3.6.0
+tcolorpy==0.1.7
+tabulate==0.10.0
+scipy==1.17.1
+portalocker==3.2.0
+pathvalidate==3.3.1
+narwhals==2.22.1
+more-itertools==11.1.0
+lxml==6.1.1
+joblib==1.5.3
+colorama==0.4.6
+chardet==6.0.0.post1
+absl-py==2.4.0
+scikit-learn==1.9.0
+sacrebleu==2.6.0
+nltk==3.9.4
+mbstrdecoder==1.1.5
+typepy==1.3.5
+rouge_score==0.1.2
+DataProperty==1.1.1
+tabledata==1.3.5
+evaluate==0.4.6
+pytablewriter==1.2.1
+lm_eval==0.4.12
+pluggy==1.6.0
+iniconfig==2.3.0
+pytest==9.1.0
+py-spy==0.4.2
+pillow==12.2.0
+tifffile==2026.6.1
+imagecodecs==2026.6.26
+zipp==4.1.0
+importlib_metadata==9.0.0
+derivative==0.6.3
+pysindy==2.1.0
+pyparsing==3.3.2
+kiwisolver==1.5.0
+fonttools==4.63.0
+cycler==0.12.1
+contourpy==1.3.3
+matplotlib==3.11.0
+sortedcontainers==2.4.0
+wcwidth==0.8.2
+sigfig==1.3.19
+importlib_resources==7.1.0
+enum-tools==0.13.0
+cramjam==2.11.0
+python-snappy==0.7.3
+compact-json==1.8.2
+numbers-parser==4.18.5

epfl_final/SparseAtt/wandb/run-20260707_170941-lgjoqt2r/files/wandb-metadata.json

@@ -0,0 +1,65 @@
+{
+  "os":  "Linux-6.8.0-124-generic-x86_64-with-glibc2.35",
+  "python":  "CPython 3.13.13",
+  "startedAt":  "2026-07-07T17:09:41.403720Z",
+  "args":  [
+    "--config",
+    "configs/ladder8k/fixed_s2.json"
+  ],
+  "program":  "/mloscratch/homes/navasard/sparse/SparseAtt/train.py",
+  "codePath":  "train.py",
+  "codePathLocal":  "train.py",
+  "git":  {
+    "remote":  "https://github.com/rubentium/TokenPrune.git",
+    "commit":  "88298c7a418622530ca5426b22b1e5ebd10ca0cb"
+  },
+  "email":  "ruben.navasardyan@mail.utoronto.ca",
+  "root":  "/mloscratch/homes/navasard/sparse/SparseAtt",
+  "host":  "h100-workspace-4-gpu-0-4",
+  "executable":  "/mloscratch/homes/navasard/conda/bin/python",
+  "cpu_count":  96,
+  "cpu_count_logical":  192,
+  "gpu":  "NVIDIA H100 80GB HBM3",
+  "gpu_count":  4,
+  "disk":  {
+    "/":  {
+      "total":  "9371709865984",
+      "used":  "5840433963008"
+    }
+  },
+  "memory":  {
+    "total":  "1623005732864"
+  },
+  "gpu_nvidia":  [
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-98029aa4-c589-b65b-3ab7-6495256f31ed"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-37a1c6c2-d3df-3064-44df-33fa82304bd0"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-b87e8111-63b6-b364-28b9-c4ce747aec08"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-bbbaaf81-9234-a1c6-0093-6d87b4e6af4c"
+    }
+  ],
+  "cudaVersion":  "13.0",
+  "writerId":  "1poap8apibhpfjc9hz6ylrlyso44bwkc"
+}

epfl_final/SparseAtt/wandb/run-20260707_170941-lgjoqt2r/files/wandb-summary.json

@@ -0,0 +1 @@
+{"train/global_step":44000,"train/loss":2.96731317281723,"eval/loss":2.9340217270851134,"eval/dense_interval_start":-1,"eval/perplexity":18.803099571771565,"eval/uses_dense_attention":0,"_wandb":{"runtime":60776},"_timestamp":1.783484544775738e+09,"_step":44000,"speed/ms_per_step":1291.2421464920044,"_runtime":60776.673024978,"train/learning_rate":0,"speed/samples_per_sec":15.48896158194279,"speed/steps_per_sec":0.7744480790971395,"eval/uses_sparse_attention":1,"eval/dense_interval_end":-1,"eval/dense_interval_active":0,"train/epoch":0.2906439083679023}

epfl_final/SparseAtt/wandb/run-20260707_170941-lgjoqt2r/logs/debug.log

@@ -0,0 +1,25 @@
+2026-07-07 17:09:41,436 INFO    MainThread:441 [wandb_setup.py:_flush():81] Current SDK version is 0.26.1
+2026-07-07 17:09:41,437 INFO    MainThread:441 [wandb_setup.py:_flush():81] Configure stats pid to 441
+2026-07-07 17:09:41,437 INFO    MainThread:441 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2026-07-07 17:09:41,437 INFO    MainThread:441 [wandb_init.py:setup_run_log_directory():723] Logging user logs to /mloscratch/homes/navasard/sparse/SparseAtt/wandb/run-20260707_170941-lgjoqt2r/logs/debug.log
+2026-07-07 17:09:41,437 INFO    MainThread:441 [wandb_init.py:setup_run_log_directory():724] Logging internal logs to /mloscratch/homes/navasard/sparse/SparseAtt/wandb/run-20260707_170941-lgjoqt2r/logs/debug-internal.log
+2026-07-07 17:09:41,437 INFO    MainThread:441 [wandb_init.py:init():850] calling init triggers
+2026-07-07 17:09:41,437 INFO    MainThread:441 [wandb_init.py:init():855] wandb.init called with sweep_config: {}
+config: {'model_path': '/mloscratch/homes/navasard/qwen3_06b_base', 'model_dtype': 'bfloat16', 'init_from_scratch': True, 'train_path': '/mloscratch/homes/navasard/data/fineweb_large/tokenized', 'val_path': None, 'max_seq_length': 8192, 'num_workers': 8, 'output_dir': '/mloscratch/homes/navasard/sparse/checkpoints/l8k365m_fixed_s2/run_20260707_034600', 'resume_from_checkpoint': '/mloscratch/homes/navasard/sparse/checkpoints/l8k365m_fixed_s2/run_20260707_034600/checkpoint-15000', 'force_new_run': False, 'num_epochs': 1, 'max_steps': 44000, 'per_device_batch_size': 2, 'gradient_accumulation_steps': 10, 'learning_rate': 0.0005, 'weight_decay': 0.01, 'warmup_steps': 750, 'lr_scheduler': 'cosine', 'max_grad_norm': 1.0, 'save_steps': 1000, 'eval_steps': 250, 'eval_batches': 50, 'logging_steps': 10, 'seed': 42, 'verbose': True, 'checkpoint_activations': False, 'strategy': 'ddp', 'mixed_precision': 'bf16', 'wandb_enabled': True, 'wandb_project': 'stochastic_sparsity', 'wandb_name': '365m fixed s2 8k', 'wandb_tags': ['qwen3', '365m', '8k', 'scaleladder', 'fixed', 'seed2'], 'hidden_size': 768, 'intermediate_size': 3072, 'num_attention_heads': 6, 'num_key_value_heads': 2, 'num_hidden_layers': 28, 'head_dim': None, 'attention_type': 'sparse_stochastic', 'sparse_block_size': 128, 'sparse_num_random_blocks': 31, 'sparse_mask_pool_size': 1, 'sparse_val_dense_compare': False, 'sparse_layer_indices': None, 'sparse_mask_cache_dir': '/mloscratch/homes/navasard/sparse/mask_cache_8k_fixed_s2', 'sparse_mask_seed': 2, 'dense_mode': False, 'dense_step_intervals': None, '_wandb': {}}
+2026-07-07 17:09:41,437 INFO    MainThread:441 [wandb_init.py:init():898] starting backend
+2026-07-07 17:09:41,983 INFO    MainThread:441 [wandb_init.py:init():913] sending inform_init request
+2026-07-07 17:09:42,429 INFO    MainThread:441 [wandb_init.py:init():918] backend started and connected
+2026-07-07 17:09:42,432 INFO    MainThread:441 [wandb_init.py:init():988] updated telemetry
+2026-07-07 17:09:42,460 INFO    MainThread:441 [wandb_init.py:init():1011] communicating run to backend with 90.0 second timeout
+2026-07-07 17:09:43,022 INFO    MainThread:441 [wandb_init.py:init():1051] run resumed
+2026-07-07 17:09:43,025 INFO    MainThread:441 [wandb_init.py:init():1056] starting run threads in backend
+2026-07-07 17:09:43,233 INFO    MainThread:441 [wandb_run.py:_console_start():2554] atexit reg
+2026-07-07 17:09:43,233 INFO    MainThread:441 [wandb_run.py:_redirect():2403] redirect: wrap_raw
+2026-07-07 17:09:43,233 INFO    MainThread:441 [wandb_run.py:_redirect():2472] Wrapping output streams.
+2026-07-07 17:09:43,233 INFO    MainThread:441 [wandb_run.py:_redirect():2495] Redirects installed.
+2026-07-07 17:09:43,241 INFO    MainThread:441 [wandb_init.py:init():1094] run started, returning control to user process
+2026-07-08 04:22:30,696 INFO    MainThread:441 [wandb_run.py:_finish():2321] finishing run ruben-navasardyan4869-epfl/stochastic_sparsity/lgjoqt2r
+2026-07-08 04:22:30,697 INFO    MainThread:441 [wandb_run.py:_atexit_cleanup():2520] got exitcode: 0
+2026-07-08 04:22:30,698 INFO    MainThread:441 [wandb_run.py:_restore():2502] restore
+2026-07-08 04:22:30,698 INFO    MainThread:441 [wandb_run.py:_restore():2508] restore done
+2026-07-08 04:22:32,673 INFO    MainThread:441 [wandb_run.py:_footer_sync_info():3889] logging synced files

epfl_final/SparseAtt/wandb/run-20260707_170941-lgjoqt2r/run-lgjoqt2r.wandb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2fde61c2d05ea9aea1e61f237354775d7fee157d4135c0bc5d5c89342e0d62c3
+size 14981098

epfl_final/SparseAtt/wandb/run-20260707_170941-qlqhc3gy/files/requirements.txt

@@ -0,0 +1,141 @@
+pip==26.0.1
+nvidia-cusparselt-cu13==0.8.0
+mpmath==1.3.0
+cuda-toolkit==13.0.2
+zstandard==0.25.0
+xxhash==3.7.0
+urllib3==2.6.3
+typing_extensions==4.15.0
+triton==3.6.0
+tqdm==4.67.3
+torchinfo==1.8.0
+sympy==1.14.0
+smmap==5.0.3
+six==1.17.0
+shellingham==1.5.4
+setuptools==81.0.0
+safetensors==0.7.0
+regex==2026.4.4
+PyYAML==6.0.3
+Pygments==2.20.0
+pyarrow==24.0.0
+psutil==7.2.2
+protobuf==7.34.1
+propcache==0.4.1
+platformdirs==4.9.6
+packaging==26.2
+nvidia-nvtx==13.0.85
+nvidia-nvshmem-cu13==3.4.5
+nvidia-nvjitlink==13.0.88
+nvidia-nccl-cu13==2.28.9
+nvidia-curand==10.4.0.35
+nvidia-cufile==1.15.1.6
+nvidia-cuda-runtime==13.0.96
+nvidia-cuda-nvrtc==13.0.88
+nvidia-cuda-cupti==13.0.85
+nvidia-cublas==13.1.0.3
+numpy==2.4.4
+networkx==3.6.1
+multidict==6.7.1
+mdurl==0.1.2
+MarkupSafe==3.0.3
+idna==3.13
+hf-xet==1.4.3
+h11==0.16.0
+fsspec==2026.2.0
+frozenlist==1.8.0
+filelock==3.29.0
+dill==0.4.1
+cuda-pathfinder==1.5.4
+click==8.3.3
+charset-normalizer==3.4.7
+certifi==2026.4.22
+attrs==26.1.0
+annotated-types==0.7.0
+annotated-doc==0.0.4
+aiohappyeyeballs==2.6.1
+yarl==1.23.0
+typing-inspection==0.4.2
+sentry-sdk==2.58.0
+requests==2.33.1
+python-dateutil==2.9.0.post0
+pydantic_core==2.46.3
+nvidia-cusparse==12.6.3.3
+nvidia-cufft==12.0.0.61
+nvidia-cudnn-cu13==9.19.0.56
+multiprocess==0.70.19
+markdown-it-py==4.0.0
+Jinja2==3.1.6
+httpcore==1.0.9
+gitdb==4.0.12
+cuda-bindings==13.2.0
+anyio==4.13.0
+aiosignal==1.4.0
+rich==15.0.0
+pydantic==2.13.3
+pandas==3.0.2
+nvidia-cusolver==12.0.4.66
+httpx==0.28.1
+GitPython==3.1.49
+aiohttp==3.13.5
+wandb==0.26.1
+typer==0.25.1
+torch==2.11.0
+huggingface_hub==1.13.0
+tokenizers==0.22.2
+datasets==4.8.5
+accelerate==1.13.0
+transformers==5.7.0
+word2number==1.1
+sqlitedict==2.1.0
+pytz==2026.2
+threadpoolctl==3.6.0
+tcolorpy==0.1.7
+tabulate==0.10.0
+scipy==1.17.1
+portalocker==3.2.0
+pathvalidate==3.3.1
+narwhals==2.22.1
+more-itertools==11.1.0
+lxml==6.1.1
+joblib==1.5.3
+colorama==0.4.6
+chardet==6.0.0.post1
+absl-py==2.4.0
+scikit-learn==1.9.0
+sacrebleu==2.6.0
+nltk==3.9.4
+mbstrdecoder==1.1.5
+typepy==1.3.5
+rouge_score==0.1.2
+DataProperty==1.1.1
+tabledata==1.3.5
+evaluate==0.4.6
+pytablewriter==1.2.1
+lm_eval==0.4.12
+pluggy==1.6.0
+iniconfig==2.3.0
+pytest==9.1.0
+py-spy==0.4.2
+pillow==12.2.0
+tifffile==2026.6.1
+imagecodecs==2026.6.26
+zipp==4.1.0
+importlib_metadata==9.0.0
+derivative==0.6.3
+pysindy==2.1.0
+pyparsing==3.3.2
+kiwisolver==1.5.0
+fonttools==4.63.0
+cycler==0.12.1
+contourpy==1.3.3
+matplotlib==3.11.0
+sortedcontainers==2.4.0
+wcwidth==0.8.2
+sigfig==1.3.19
+importlib_resources==7.1.0
+enum-tools==0.13.0
+cramjam==2.11.0
+python-snappy==0.7.3
+compact-json==1.8.2
+numbers-parser==4.18.5

epfl_final/SparseAtt/wandb/run-20260707_170941-qlqhc3gy/run-qlqhc3gy.wandb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:39f0b4c92db0cce6424349ea82089ea6af47d1fbe0f9cfef25664a8a2aa06916
+size 15466496

epfl_final/SparseAtt/wandb/run-20260707_171131-304r0ilz/files/config.yaml

@@ -0,0 +1,188 @@
+_wandb:
+    value:
+        cli_version: 0.26.1
+        e:
+            aocwb8ws6r6lkksmxs5nsgnib0p4xppe:
+                args:
+                    - --config
+                    - configs/ladder8k/resampled.json
+                codePath: train.py
+                codePathLocal: train.py
+                cpu_count: 96
+                cpu_count_logical: 192
+                cudaVersion: "13.0"
+                disk:
+                    /:
+                        total: "9371709865984"
+                        used: "5840501911552"
+                email: ruben.navasardyan@mail.utoronto.ca
+                executable: /mloscratch/homes/navasard/conda/bin/python
+                git:
+                    commit: 88298c7a418622530ca5426b22b1e5ebd10ca0cb
+                    remote: https://github.com/rubentium/TokenPrune.git
+                gpu: NVIDIA H100 80GB HBM3
+                gpu_count: 4
+                gpu_nvidia:
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-98029aa4-c589-b65b-3ab7-6495256f31ed
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-37a1c6c2-d3df-3064-44df-33fa82304bd0
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-b87e8111-63b6-b364-28b9-c4ce747aec08
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-bbbaaf81-9234-a1c6-0093-6d87b4e6af4c
+                host: h100-workspace-4-gpu-0-4
+                memory:
+                    total: "1623005732864"
+                os: Linux-6.8.0-124-generic-x86_64-with-glibc2.35
+                program: /mloscratch/homes/navasard/sparse/SparseAtt/train.py
+                python: CPython 3.13.13
+                root: /mloscratch/homes/navasard/sparse/SparseAtt
+                startedAt: "2026-07-07T17:11:31.714252Z"
+                writerId: aocwb8ws6r6lkksmxs5nsgnib0p4xppe
+        m: []
+        python_version: 3.13.13
+        t:
+            "1":
+                - 1
+                - 5
+                - 11
+                - 49
+                - 51
+                - 53
+                - 71
+            "2":
+                - 1
+                - 5
+                - 11
+                - 49
+                - 51
+                - 53
+                - 71
+            "3":
+                - 2
+                - 5
+                - 13
+                - 14
+                - 15
+                - 16
+                - 61
+                - 62
+            "4": 3.13.13
+            "5": 0.26.1
+            "6": 5.7.0
+            "12": 0.26.1
+            "13": linux-x86_64
+attention_type:
+    value: sparse_stochastic
+checkpoint_activations:
+    value: false
+dense_mode:
+    value: false
+dense_step_intervals:
+    value: null
+eval_batches:
+    value: 50
+eval_steps:
+    value: 250
+force_new_run:
+    value: false
+gradient_accumulation_steps:
+    value: 10
+head_dim:
+    value: null
+hidden_size:
+    value: 768
+init_from_scratch:
+    value: true
+intermediate_size:
+    value: 3072
+learning_rate:
+    value: 0.0005
+logging_steps:
+    value: 10
+lr_scheduler:
+    value: cosine
+max_grad_norm:
+    value: 1
+max_seq_length:
+    value: 8192
+max_steps:
+    value: 44000
+mixed_precision:
+    value: bf16
+model_dtype:
+    value: bfloat16
+model_path:
+    value: /mloscratch/homes/navasard/qwen3_06b_base
+num_attention_heads:
+    value: 6
+num_epochs:
+    value: 1
+num_hidden_layers:
+    value: 28
+num_key_value_heads:
+    value: 2
+num_workers:
+    value: 8
+output_dir:
+    value: /mloscratch/homes/navasard/sparse/checkpoints/l8k365m_resampled/run_20260707_034554
+per_device_batch_size:
+    value: 2
+resume_from_checkpoint:
+    value: /mloscratch/homes/navasard/sparse/checkpoints/l8k365m_resampled/run_20260707_034554/checkpoint-14000
+save_steps:
+    value: 1000
+seed:
+    value: 42
+sparse_block_size:
+    value: 128
+sparse_layer_indices:
+    value: null
+sparse_mask_cache_dir:
+    value: /mloscratch/homes/navasard/sparse/mask_cache_32
+sparse_mask_pool_size:
+    value: 10000
+sparse_mask_seed:
+    value: 0
+sparse_num_random_blocks:
+    value: 31
+sparse_val_dense_compare:
+    value: false
+strategy:
+    value: ddp
+train_path:
+    value: /mloscratch/homes/navasard/data/fineweb_large/tokenized
+val_path:
+    value: null
+verbose:
+    value: true
+wandb_enabled:
+    value: true
+wandb_name:
+    value: 365m resampled 8k nrb31
+wandb_project:
+    value: stochastic_sparsity
+wandb_tags:
+    value:
+        - qwen3
+        - 365m
+        - 8k
+        - scaleladder
+        - resampled
+warmup_steps:
+    value: 750
+weight_decay:
+    value: 0.01

epfl_final/SparseAtt/wandb/run-20260707_171131-304r0ilz/files/requirements.txt

@@ -0,0 +1,141 @@
+pip==26.0.1
+nvidia-cusparselt-cu13==0.8.0
+mpmath==1.3.0
+cuda-toolkit==13.0.2
+zstandard==0.25.0
+xxhash==3.7.0
+urllib3==2.6.3
+typing_extensions==4.15.0
+triton==3.6.0
+tqdm==4.67.3
+torchinfo==1.8.0
+sympy==1.14.0
+smmap==5.0.3
+six==1.17.0
+shellingham==1.5.4
+setuptools==81.0.0
+safetensors==0.7.0
+regex==2026.4.4
+PyYAML==6.0.3
+Pygments==2.20.0
+pyarrow==24.0.0
+psutil==7.2.2
+protobuf==7.34.1
+propcache==0.4.1
+platformdirs==4.9.6
+packaging==26.2
+nvidia-nvtx==13.0.85
+nvidia-nvshmem-cu13==3.4.5
+nvidia-nvjitlink==13.0.88
+nvidia-nccl-cu13==2.28.9
+nvidia-curand==10.4.0.35
+nvidia-cufile==1.15.1.6
+nvidia-cuda-runtime==13.0.96
+nvidia-cuda-nvrtc==13.0.88
+nvidia-cuda-cupti==13.0.85
+nvidia-cublas==13.1.0.3
+numpy==2.4.4
+networkx==3.6.1
+multidict==6.7.1
+mdurl==0.1.2
+MarkupSafe==3.0.3
+idna==3.13
+hf-xet==1.4.3
+h11==0.16.0
+fsspec==2026.2.0
+frozenlist==1.8.0
+filelock==3.29.0
+dill==0.4.1
+cuda-pathfinder==1.5.4
+click==8.3.3
+charset-normalizer==3.4.7
+certifi==2026.4.22
+attrs==26.1.0
+annotated-types==0.7.0
+annotated-doc==0.0.4
+aiohappyeyeballs==2.6.1
+yarl==1.23.0
+typing-inspection==0.4.2
+sentry-sdk==2.58.0
+requests==2.33.1
+python-dateutil==2.9.0.post0
+pydantic_core==2.46.3
+nvidia-cusparse==12.6.3.3
+nvidia-cufft==12.0.0.61
+nvidia-cudnn-cu13==9.19.0.56
+multiprocess==0.70.19
+markdown-it-py==4.0.0
+Jinja2==3.1.6
+httpcore==1.0.9
+gitdb==4.0.12
+cuda-bindings==13.2.0
+anyio==4.13.0
+aiosignal==1.4.0
+rich==15.0.0
+pydantic==2.13.3
+pandas==3.0.2
+nvidia-cusolver==12.0.4.66
+httpx==0.28.1
+GitPython==3.1.49
+aiohttp==3.13.5
+wandb==0.26.1
+typer==0.25.1
+torch==2.11.0
+huggingface_hub==1.13.0
+tokenizers==0.22.2
+datasets==4.8.5
+accelerate==1.13.0
+transformers==5.7.0
+word2number==1.1
+sqlitedict==2.1.0
+pytz==2026.2
+threadpoolctl==3.6.0
+tcolorpy==0.1.7
+tabulate==0.10.0
+scipy==1.17.1
+portalocker==3.2.0
+pathvalidate==3.3.1
+narwhals==2.22.1
+more-itertools==11.1.0
+lxml==6.1.1
+joblib==1.5.3
+colorama==0.4.6
+chardet==6.0.0.post1
+absl-py==2.4.0
+scikit-learn==1.9.0
+sacrebleu==2.6.0
+nltk==3.9.4
+mbstrdecoder==1.1.5
+typepy==1.3.5
+rouge_score==0.1.2
+DataProperty==1.1.1
+tabledata==1.3.5
+evaluate==0.4.6
+pytablewriter==1.2.1
+lm_eval==0.4.12
+pluggy==1.6.0
+iniconfig==2.3.0
+pytest==9.1.0
+py-spy==0.4.2
+pillow==12.2.0
+tifffile==2026.6.1
+imagecodecs==2026.6.26
+zipp==4.1.0
+importlib_metadata==9.0.0
+derivative==0.6.3
+pysindy==2.1.0
+pyparsing==3.3.2
+kiwisolver==1.5.0
+fonttools==4.63.0
+cycler==0.12.1
+contourpy==1.3.3
+matplotlib==3.11.0
+sortedcontainers==2.4.0
+wcwidth==0.8.2
+sigfig==1.3.19
+importlib_resources==7.1.0
+enum-tools==0.13.0
+cramjam==2.11.0
+python-snappy==0.7.3
+compact-json==1.8.2
+numbers-parser==4.18.5

epfl_final/SparseAtt/wandb/run-20260707_171131-304r0ilz/files/wandb-metadata.json

@@ -0,0 +1,65 @@
+{
+  "os":  "Linux-6.8.0-124-generic-x86_64-with-glibc2.35",
+  "python":  "CPython 3.13.13",
+  "startedAt":  "2026-07-07T17:11:31.714252Z",
+  "args":  [
+    "--config",
+    "configs/ladder8k/resampled.json"
+  ],
+  "program":  "/mloscratch/homes/navasard/sparse/SparseAtt/train.py",
+  "codePath":  "train.py",
+  "codePathLocal":  "train.py",
+  "git":  {
+    "remote":  "https://github.com/rubentium/TokenPrune.git",
+    "commit":  "88298c7a418622530ca5426b22b1e5ebd10ca0cb"
+  },
+  "email":  "ruben.navasardyan@mail.utoronto.ca",
+  "root":  "/mloscratch/homes/navasard/sparse/SparseAtt",
+  "host":  "h100-workspace-4-gpu-0-4",
+  "executable":  "/mloscratch/homes/navasard/conda/bin/python",
+  "cpu_count":  96,
+  "cpu_count_logical":  192,
+  "gpu":  "NVIDIA H100 80GB HBM3",
+  "gpu_count":  4,
+  "disk":  {
+    "/":  {
+      "total":  "9371709865984",
+      "used":  "5840501911552"
+    }
+  },
+  "memory":  {
+    "total":  "1623005732864"
+  },
+  "gpu_nvidia":  [
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-98029aa4-c589-b65b-3ab7-6495256f31ed"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-37a1c6c2-d3df-3064-44df-33fa82304bd0"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-b87e8111-63b6-b364-28b9-c4ce747aec08"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-bbbaaf81-9234-a1c6-0093-6d87b4e6af4c"
+    }
+  ],
+  "cudaVersion":  "13.0",
+  "writerId":  "aocwb8ws6r6lkksmxs5nsgnib0p4xppe"
+}

epfl_final/SparseAtt/wandb/run-20260707_171131-304r0ilz/files/wandb-summary.json

@@ -0,0 +1 @@
+{"train/global_step":44000,"speed/ms_per_step":1279.413390159607,"train/learning_rate":0,"_step":44000,"speed/samples_per_sec":15.632164047857119,"_wandb":{"runtime":61700},"_runtime":61700.494997317,"_timestamp":1.7834856746005292e+09,"eval/dense_interval_active":0,"speed/steps_per_sec":0.7816082023928559,"eval/dense_interval_end":-1,"eval/dense_interval_start":-1,"train/epoch":0.2906439083679023,"eval/perplexity":18.88360197859757,"eval/uses_dense_attention":0,"train/loss":2.9724996876716614,"eval/loss":2.9382939252853393,"eval/uses_sparse_attention":1}

epfl_final/SparseAtt/wandb/run-20260707_171131-304r0ilz/logs/debug.log

@@ -0,0 +1,25 @@
+2026-07-07 17:11:31,740 INFO    MainThread:420 [wandb_setup.py:_flush():81] Current SDK version is 0.26.1
+2026-07-07 17:11:31,740 INFO    MainThread:420 [wandb_setup.py:_flush():81] Configure stats pid to 420
+2026-07-07 17:11:31,740 INFO    MainThread:420 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2026-07-07 17:11:31,740 INFO    MainThread:420 [wandb_init.py:setup_run_log_directory():723] Logging user logs to /mloscratch/homes/navasard/sparse/SparseAtt/wandb/run-20260707_171131-304r0ilz/logs/debug.log
+2026-07-07 17:11:31,740 INFO    MainThread:420 [wandb_init.py:setup_run_log_directory():724] Logging internal logs to /mloscratch/homes/navasard/sparse/SparseAtt/wandb/run-20260707_171131-304r0ilz/logs/debug-internal.log
+2026-07-07 17:11:31,740 INFO    MainThread:420 [wandb_init.py:init():850] calling init triggers
+2026-07-07 17:11:31,740 INFO    MainThread:420 [wandb_init.py:init():855] wandb.init called with sweep_config: {}
+config: {'model_path': '/mloscratch/homes/navasard/qwen3_06b_base', 'model_dtype': 'bfloat16', 'init_from_scratch': True, 'train_path': '/mloscratch/homes/navasard/data/fineweb_large/tokenized', 'val_path': None, 'max_seq_length': 8192, 'num_workers': 8, 'output_dir': '/mloscratch/homes/navasard/sparse/checkpoints/l8k365m_resampled/run_20260707_034554', 'resume_from_checkpoint': '/mloscratch/homes/navasard/sparse/checkpoints/l8k365m_resampled/run_20260707_034554/checkpoint-14000', 'force_new_run': False, 'num_epochs': 1, 'max_steps': 44000, 'per_device_batch_size': 2, 'gradient_accumulation_steps': 10, 'learning_rate': 0.0005, 'weight_decay': 0.01, 'warmup_steps': 750, 'lr_scheduler': 'cosine', 'max_grad_norm': 1.0, 'save_steps': 1000, 'eval_steps': 250, 'eval_batches': 50, 'logging_steps': 10, 'seed': 42, 'verbose': True, 'checkpoint_activations': False, 'strategy': 'ddp', 'mixed_precision': 'bf16', 'wandb_enabled': True, 'wandb_project': 'stochastic_sparsity', 'wandb_name': '365m resampled 8k nrb31', 'wandb_tags': ['qwen3', '365m', '8k', 'scaleladder', 'resampled'], 'hidden_size': 768, 'intermediate_size': 3072, 'num_attention_heads': 6, 'num_key_value_heads': 2, 'num_hidden_layers': 28, 'head_dim': None, 'attention_type': 'sparse_stochastic', 'sparse_block_size': 128, 'sparse_num_random_blocks': 31, 'sparse_mask_pool_size': 10000, 'sparse_val_dense_compare': False, 'sparse_layer_indices': None, 'sparse_mask_cache_dir': '/mloscratch/homes/navasard/sparse/mask_cache_32', 'sparse_mask_seed': 0, 'dense_mode': False, 'dense_step_intervals': None, '_wandb': {}}
+2026-07-07 17:11:31,740 INFO    MainThread:420 [wandb_init.py:init():898] starting backend
+2026-07-07 17:11:32,008 INFO    MainThread:420 [wandb_init.py:init():913] sending inform_init request
+2026-07-07 17:11:32,376 INFO    MainThread:420 [wandb_init.py:init():918] backend started and connected
+2026-07-07 17:11:32,378 INFO    MainThread:420 [wandb_init.py:init():988] updated telemetry
+2026-07-07 17:11:32,396 INFO    MainThread:420 [wandb_init.py:init():1011] communicating run to backend with 90.0 second timeout
+2026-07-07 17:11:33,015 INFO    MainThread:420 [wandb_init.py:init():1051] run resumed
+2026-07-07 17:11:33,017 INFO    MainThread:420 [wandb_init.py:init():1056] starting run threads in backend
+2026-07-07 17:11:33,296 INFO    MainThread:420 [wandb_run.py:_console_start():2554] atexit reg
+2026-07-07 17:11:33,296 INFO    MainThread:420 [wandb_run.py:_redirect():2403] redirect: wrap_raw
+2026-07-07 17:11:33,296 INFO    MainThread:420 [wandb_run.py:_redirect():2472] Wrapping output streams.
+2026-07-07 17:11:33,296 INFO    MainThread:420 [wandb_run.py:_redirect():2495] Redirects installed.
+2026-07-07 17:11:33,302 INFO    MainThread:420 [wandb_init.py:init():1094] run started, returning control to user process
+2026-07-08 04:41:20,512 INFO    MainThread:420 [wandb_run.py:_finish():2321] finishing run ruben-navasardyan4869-epfl/stochastic_sparsity/304r0ilz
+2026-07-08 04:41:20,513 INFO    MainThread:420 [wandb_run.py:_atexit_cleanup():2520] got exitcode: 0
+2026-07-08 04:41:20,513 INFO    MainThread:420 [wandb_run.py:_restore():2502] restore
+2026-07-08 04:41:20,513 INFO    MainThread:420 [wandb_run.py:_restore():2508] restore done
+2026-07-08 04:41:21,979 INFO    MainThread:420 [wandb_run.py:_footer_sync_info():3889] logging synced files

epfl_final/SparseAtt/wandb/run-20260707_171131-304r0ilz/run-304r0ilz.wandb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bcc2351b898ff2a8faebbabbdbd19f0c98e171fe43317be04bfa0500ca90f7f6
+size 15331216

epfl_final/SparseAtt/wandb/run-20260708_045920-qlqhc3gy/files/output.log

@@ -0,0 +1,235 @@
+Starting training...
+Validation at step 41000: Loss = 2.8935, Perplexity = 18.06
+wandb: WARNING Tried to log to step 41000 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Switching attention mode for step 41001: dense (dense step interval)
+Step 41010/44000 | Loss: 2.8830 | LR: 5.87e-06 | 72705.1ms/step | 0.3 samples/s
+wandb: WARNING Tried to log to step 41010 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41020/44000 | Loss: 2.8841 | LR: 5.83e-06 | 1378.2ms/step | 14.5 samples/s
+wandb: WARNING Tried to log to step 41020 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41030/44000 | Loss: 2.9077 | LR: 5.80e-06 | 1378.7ms/step | 14.5 samples/s
+wandb: WARNING Tried to log to step 41030 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41040/44000 | Loss: 2.8674 | LR: 5.76e-06 | 1382.7ms/step | 14.5 samples/s
+wandb: WARNING Tried to log to step 41040 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41050/44000 | Loss: 2.8637 | LR: 5.72e-06 | 1384.4ms/step | 14.4 samples/s
+wandb: WARNING Tried to log to step 41050 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41060/44000 | Loss: 2.8630 | LR: 5.68e-06 | 1382.8ms/step | 14.5 samples/s
+wandb: WARNING Tried to log to step 41060 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41070/44000 | Loss: 2.8783 | LR: 5.64e-06 | 1387.2ms/step | 14.4 samples/s
+wandb: WARNING Tried to log to step 41070 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41080/44000 | Loss: 2.8739 | LR: 5.60e-06 | 1384.3ms/step | 14.4 samples/s
+wandb: WARNING Tried to log to step 41080 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41090/44000 | Loss: 2.8919 | LR: 5.56e-06 | 1385.1ms/step | 14.4 samples/s
+wandb: WARNING Tried to log to step 41090 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41100/44000 | Loss: 2.8743 | LR: 5.53e-06 | 1384.7ms/step | 14.4 samples/s
+wandb: WARNING Tried to log to step 41100 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41110/44000 | Loss: 2.8836 | LR: 5.49e-06 | 1388.0ms/step | 14.4 samples/s
+wandb: WARNING Tried to log to step 41110 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41120/44000 | Loss: 2.8531 | LR: 5.45e-06 | 1388.0ms/step | 14.4 samples/s
+wandb: WARNING Tried to log to step 41120 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41130/44000 | Loss: 2.8677 | LR: 5.41e-06 | 1386.2ms/step | 14.4 samples/s
+wandb: WARNING Tried to log to step 41130 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41140/44000 | Loss: 2.8631 | LR: 5.38e-06 | 1385.5ms/step | 14.4 samples/s
+wandb: WARNING Tried to log to step 41140 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41150/44000 | Loss: 2.8616 | LR: 5.34e-06 | 1388.3ms/step | 14.4 samples/s
+wandb: WARNING Tried to log to step 41150 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41160/44000 | Loss: 2.8721 | LR: 5.30e-06 | 1388.0ms/step | 14.4 samples/s
+wandb: WARNING Tried to log to step 41160 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41170/44000 | Loss: 2.8751 | LR: 5.26e-06 | 1386.8ms/step | 14.4 samples/s
+wandb: WARNING Tried to log to step 41170 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41180/44000 | Loss: 2.8789 | LR: 5.23e-06 | 1387.4ms/step | 14.4 samples/s
+wandb: WARNING Tried to log to step 41180 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41190/44000 | Loss: 2.8825 | LR: 5.19e-06 | 1387.6ms/step | 14.4 samples/s
+wandb: WARNING Tried to log to step 41190 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41200/44000 | Loss: 2.8272 | LR: 5.15e-06 | 1387.6ms/step | 14.4 samples/s
+wandb: WARNING Tried to log to step 41200 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41210/44000 | Loss: 2.8894 | LR: 5.12e-06 | 1388.4ms/step | 14.4 samples/s
+wandb: WARNING Tried to log to step 41210 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41220/44000 | Loss: 2.8786 | LR: 5.08e-06 | 1388.1ms/step | 14.4 samples/s
+wandb: WARNING Tried to log to step 41220 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41230/44000 | Loss: 2.8999 | LR: 5.04e-06 | 1385.5ms/step | 14.4 samples/s
+wandb: WARNING Tried to log to step 41230 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41240/44000 | Loss: 2.8691 | LR: 5.01e-06 | 1386.3ms/step | 14.4 samples/s
+wandb: WARNING Tried to log to step 41240 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41250/44000 | Loss: 2.8575 | LR: 4.97e-06 | 1386.4ms/step | 14.4 samples/s
+wandb: WARNING Tried to log to step 41250 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Validation at step 41250: Loss = 2.8964, Perplexity = 18.11
+Step 41260/44000 | Loss: 2.8636 | LR: 4.94e-06 | 3599.9ms/step | 5.6 samples/s
+wandb: WARNING Tried to log to step 41260 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41270/44000 | Loss: 2.8639 | LR: 4.90e-06 | 1395.9ms/step | 14.3 samples/s
+wandb: WARNING Tried to log to step 41270 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41280/44000 | Loss: 2.8776 | LR: 4.86e-06 | 1397.4ms/step | 14.3 samples/s
+wandb: WARNING Tried to log to step 41280 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41290/44000 | Loss: 2.8910 | LR: 4.83e-06 | 1392.4ms/step | 14.4 samples/s
+wandb: WARNING Tried to log to step 41290 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41300/44000 | Loss: 2.8666 | LR: 4.79e-06 | 1394.1ms/step | 14.3 samples/s
+wandb: WARNING Tried to log to step 41300 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41310/44000 | Loss: 2.8688 | LR: 4.76e-06 | 1394.6ms/step | 14.3 samples/s
+wandb: WARNING Tried to log to step 41310 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41320/44000 | Loss: 2.8846 | LR: 4.72e-06 | 1395.5ms/step | 14.3 samples/s
+wandb: WARNING Tried to log to step 41320 that is less than the current step 41321. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 41330/44000 | Loss: 2.8830 | LR: 4.69e-06 | 1398.5ms/step | 14.3 samples/s
+Step 41340/44000 | Loss: 2.8774 | LR: 4.65e-06 | 1399.8ms/step | 14.3 samples/s
+Step 41350/44000 | Loss: 2.8814 | LR: 4.62e-06 | 1399.9ms/step | 14.3 samples/s
+Step 41360/44000 | Loss: 2.8783 | LR: 4.58e-06 | 1397.8ms/step | 14.3 samples/s
+Step 41370/44000 | Loss: 2.8751 | LR: 4.55e-06 | 1399.9ms/step | 14.3 samples/s
+Step 41380/44000 | Loss: 2.8753 | LR: 4.51e-06 | 1400.2ms/step | 14.3 samples/s
+Step 41390/44000 | Loss: 2.8803 | LR: 4.48e-06 | 1398.1ms/step | 14.3 samples/s
+Step 41400/44000 | Loss: 2.8622 | LR: 4.45e-06 | 1399.9ms/step | 14.3 samples/s
+Step 41410/44000 | Loss: 2.8952 | LR: 4.41e-06 | 1400.0ms/step | 14.3 samples/s
+Step 41420/44000 | Loss: 2.9003 | LR: 4.38e-06 | 1400.0ms/step | 14.3 samples/s
+Step 41430/44000 | Loss: 2.8942 | LR: 4.34e-06 | 1400.1ms/step | 14.3 samples/s
+Step 41440/44000 | Loss: 2.8893 | LR: 4.31e-06 | 1400.2ms/step | 14.3 samples/s
+Step 41450/44000 | Loss: 2.8893 | LR: 4.28e-06 | 1397.7ms/step | 14.3 samples/s
+Step 41460/44000 | Loss: 2.8877 | LR: 4.24e-06 | 1399.8ms/step | 14.3 samples/s
+Step 41470/44000 | Loss: 2.8502 | LR: 4.21e-06 | 1399.9ms/step | 14.3 samples/s
+Step 41480/44000 | Loss: 2.9006 | LR: 4.18e-06 | 1398.2ms/step | 14.3 samples/s
+Step 41490/44000 | Loss: 2.8452 | LR: 4.14e-06 | 1398.0ms/step | 14.3 samples/s
+Step 41500/44000 | Loss: 2.8632 | LR: 4.11e-06 | 1395.5ms/step | 14.3 samples/s
+Validation at step 41500: Loss = 2.8957, Perplexity = 18.10
+Step 41510/44000 | Loss: 2.8852 | LR: 4.08e-06 | 3608.6ms/step | 5.5 samples/s
+Step 41520/44000 | Loss: 2.8848 | LR: 4.05e-06 | 1400.0ms/step | 14.3 samples/s
+Step 41530/44000 | Loss: 2.8956 | LR: 4.01e-06 | 1399.9ms/step | 14.3 samples/s
+Step 41540/44000 | Loss: 2.8755 | LR: 3.98e-06 | 1398.0ms/step | 14.3 samples/s
+Step 41550/44000 | Loss: 2.8957 | LR: 3.95e-06 | 1397.9ms/step | 14.3 samples/s
+Step 41560/44000 | Loss: 2.8824 | LR: 3.92e-06 | 1398.5ms/step | 14.3 samples/s
+Step 41570/44000 | Loss: 2.8392 | LR: 3.88e-06 | 1399.5ms/step | 14.3 samples/s
+Step 41580/44000 | Loss: 2.8638 | LR: 3.85e-06 | 1399.9ms/step | 14.3 samples/s
+Step 41590/44000 | Loss: 2.9169 | LR: 3.82e-06 | 1399.9ms/step | 14.3 samples/s
+Step 41600/44000 | Loss: 2.8774 | LR: 3.79e-06 | 1400.3ms/step | 14.3 samples/s
+Step 41610/44000 | Loss: 2.8550 | LR: 3.76e-06 | 1400.2ms/step | 14.3 samples/s
+Step 41620/44000 | Loss: 2.8866 | LR: 3.73e-06 | 1397.6ms/step | 14.3 samples/s
+Step 41630/44000 | Loss: 2.9064 | LR: 3.70e-06 | 1398.2ms/step | 14.3 samples/s
+Step 41640/44000 | Loss: 2.8708 | LR: 3.66e-06 | 1395.8ms/step | 14.3 samples/s
+Step 41650/44000 | Loss: 2.8892 | LR: 3.63e-06 | 1396.9ms/step | 14.3 samples/s
+Step 41660/44000 | Loss: 2.8436 | LR: 3.60e-06 | 1399.3ms/step | 14.3 samples/s
+Step 41670/44000 | Loss: 2.8685 | LR: 3.57e-06 | 1396.1ms/step | 14.3 samples/s
+Step 41680/44000 | Loss: 2.9064 | LR: 3.54e-06 | 1398.6ms/step | 14.3 samples/s
+Step 41690/44000 | Loss: 2.8932 | LR: 3.51e-06 | 1399.6ms/step | 14.3 samples/s
+Step 41700/44000 | Loss: 2.8826 | LR: 3.48e-06 | 1397.9ms/step | 14.3 samples/s
+Step 41710/44000 | Loss: 2.8852 | LR: 3.45e-06 | 1397.6ms/step | 14.3 samples/s
+Step 41720/44000 | Loss: 2.8716 | LR: 3.42e-06 | 1398.7ms/step | 14.3 samples/s
+Step 41730/44000 | Loss: 2.8885 | LR: 3.39e-06 | 1396.9ms/step | 14.3 samples/s
+Step 41740/44000 | Loss: 2.8730 | LR: 3.36e-06 | 1394.7ms/step | 14.3 samples/s
+Step 41750/44000 | Loss: 2.8644 | LR: 3.33e-06 | 1396.2ms/step | 14.3 samples/s
+Validation at step 41750: Loss = 2.8955, Perplexity = 18.09
+Step 41760/44000 | Loss: 2.8922 | LR: 3.30e-06 | 3605.9ms/step | 5.5 samples/s
+Step 41770/44000 | Loss: 2.8848 | LR: 3.27e-06 | 1400.2ms/step | 14.3 samples/s
+Step 41780/44000 | Loss: 2.8771 | LR: 3.24e-06 | 1399.9ms/step | 14.3 samples/s
+Step 41790/44000 | Loss: 2.8795 | LR: 3.21e-06 | 1398.4ms/step | 14.3 samples/s
+Step 41800/44000 | Loss: 2.9094 | LR: 3.19e-06 | 1399.4ms/step | 14.3 samples/s
+Step 41810/44000 | Loss: 2.8947 | LR: 3.16e-06 | 1400.0ms/step | 14.3 samples/s
+Step 41820/44000 | Loss: 2.8932 | LR: 3.13e-06 | 1398.6ms/step | 14.3 samples/s
+Step 41830/44000 | Loss: 2.8873 | LR: 3.10e-06 | 1399.7ms/step | 14.3 samples/s
+Step 41840/44000 | Loss: 2.8630 | LR: 3.07e-06 | 1398.1ms/step | 14.3 samples/s
+Step 41850/44000 | Loss: 2.8938 | LR: 3.04e-06 | 1399.7ms/step | 14.3 samples/s
+Step 41860/44000 | Loss: 2.8598 | LR: 3.01e-06 | 1398.1ms/step | 14.3 samples/s
+Step 41870/44000 | Loss: 2.8600 | LR: 2.99e-06 | 1397.9ms/step | 14.3 samples/s
+Step 41880/44000 | Loss: 2.8726 | LR: 2.96e-06 | 1399.5ms/step | 14.3 samples/s
+Step 41890/44000 | Loss: 2.8900 | LR: 2.93e-06 | 1396.4ms/step | 14.3 samples/s
+Step 41900/44000 | Loss: 2.8912 | LR: 2.90e-06 | 1396.2ms/step | 14.3 samples/s
+Step 41910/44000 | Loss: 2.8584 | LR: 2.88e-06 | 1398.3ms/step | 14.3 samples/s
+Step 41920/44000 | Loss: 2.8649 | LR: 2.85e-06 | 1399.6ms/step | 14.3 samples/s
+Step 41930/44000 | Loss: 2.8647 | LR: 2.82e-06 | 1398.6ms/step | 14.3 samples/s
+Step 41940/44000 | Loss: 2.8775 | LR: 2.79e-06 | 1399.8ms/step | 14.3 samples/s
+Step 41950/44000 | Loss: 2.8947 | LR: 2.77e-06 | 1399.9ms/step | 14.3 samples/s
+Step 41960/44000 | Loss: 2.8844 | LR: 2.74e-06 | 1399.8ms/step | 14.3 samples/s
+Step 41970/44000 | Loss: 2.8521 | LR: 2.71e-06 | 1400.3ms/step | 14.3 samples/s
+Step 41980/44000 | Loss: 2.8934 | LR: 2.69e-06 | 1399.7ms/step | 14.3 samples/s
+Step 41990/44000 | Loss: 2.8872 | LR: 2.66e-06 | 1400.2ms/step | 14.3 samples/s
+Step 42000/44000 | Loss: 2.8674 | LR: 2.63e-06 | 1399.8ms/step | 14.3 samples/s
+Validation at step 42000: Loss = 2.8955, Perplexity = 18.09
+Saved checkpoint to /mloscratch/homes/navasard/sparse/checkpoints/l8k365m_dense/run_20260707_034553/checkpoint-42000
+Step 42010/44000 | Loss: 2.8728 | LR: 2.61e-06 | 3947.4ms/step | 5.1 samples/s
+Step 42020/44000 | Loss: 2.8740 | LR: 2.58e-06 | 1387.7ms/step | 14.4 samples/s
+Step 42030/44000 | Loss: 2.9064 | LR: 2.56e-06 | 1386.0ms/step | 14.4 samples/s
+Step 42040/44000 | Loss: 2.8997 | LR: 2.53e-06 | 1387.4ms/step | 14.4 samples/s
+Step 42050/44000 | Loss: 2.9047 | LR: 2.50e-06 | 1387.6ms/step | 14.4 samples/s
+Step 42060/44000 | Loss: 2.8732 | LR: 2.48e-06 | 1387.6ms/step | 14.4 samples/s
+Step 42070/44000 | Loss: 2.8942 | LR: 2.45e-06 | 1388.5ms/step | 14.4 samples/s
+Step 42080/44000 | Loss: 2.8655 | LR: 2.43e-06 | 1400.2ms/step | 14.3 samples/s
+Step 42090/44000 | Loss: 2.8531 | LR: 2.40e-06 | 1392.0ms/step | 14.4 samples/s
+Step 42100/44000 | Loss: 2.8703 | LR: 2.38e-06 | 1383.8ms/step | 14.5 samples/s
+Step 42110/44000 | Loss: 2.8540 | LR: 2.35e-06 | 1390.4ms/step | 14.4 samples/s
+Step 42120/44000 | Loss: 2.8734 | LR: 2.33e-06 | 1389.7ms/step | 14.4 samples/s
+Step 42130/44000 | Loss: 2.8712 | LR: 2.30e-06 | 1385.4ms/step | 14.4 samples/s
+Step 42140/44000 | Loss: 2.8697 | LR: 2.28e-06 | 1384.9ms/step | 14.4 samples/s
+Step 42150/44000 | Loss: 2.8552 | LR: 2.25e-06 | 1385.8ms/step | 14.4 samples/s
+Step 42160/44000 | Loss: 2.8535 | LR: 2.23e-06 | 1389.3ms/step | 14.4 samples/s
+Step 42170/44000 | Loss: 2.8419 | LR: 2.21e-06 | 1386.3ms/step | 14.4 samples/s
+Step 42180/44000 | Loss: 2.8625 | LR: 2.18e-06 | 1383.7ms/step | 14.5 samples/s
+Step 42190/44000 | Loss: 2.8945 | LR: 2.16e-06 | 1384.8ms/step | 14.4 samples/s
+Step 42200/44000 | Loss: 2.8659 | LR: 2.13e-06 | 1391.9ms/step | 14.4 samples/s
+Step 42210/44000 | Loss: 2.8400 | LR: 2.11e-06 | 1394.3ms/step | 14.3 samples/s
+Step 42220/44000 | Loss: 2.8939 | LR: 2.09e-06 | 1389.2ms/step | 14.4 samples/s
+Step 42230/44000 | Loss: 2.8788 | LR: 2.06e-06 | 1386.2ms/step | 14.4 samples/s
+Step 42240/44000 | Loss: 2.9296 | LR: 2.04e-06 | 1384.1ms/step | 14.4 samples/s
+Step 42250/44000 | Loss: 2.8619 | LR: 2.02e-06 | 1386.7ms/step | 14.4 samples/s
+Validation at step 42250: Loss = 2.8954, Perplexity = 18.09
+Step 42260/44000 | Loss: 2.8622 | LR: 1.99e-06 | 3581.7ms/step | 5.6 samples/s
+Step 42270/44000 | Loss: 2.8559 | LR: 1.97e-06 | 1387.6ms/step | 14.4 samples/s
+Step 42280/44000 | Loss: 2.8763 | LR: 1.95e-06 | 1384.4ms/step | 14.4 samples/s
+Step 42290/44000 | Loss: 2.8733 | LR: 1.93e-06 | 1386.8ms/step | 14.4 samples/s
+Step 42300/44000 | Loss: 2.8489 | LR: 1.90e-06 | 1388.0ms/step | 14.4 samples/s
+Step 42310/44000 | Loss: 2.8729 | LR: 1.88e-06 | 1384.8ms/step | 14.4 samples/s
+Step 42320/44000 | Loss: 2.8861 | LR: 1.86e-06 | 1384.6ms/step | 14.4 samples/s
+Step 42330/44000 | Loss: 2.8642 | LR: 1.84e-06 | 1387.6ms/step | 14.4 samples/s
+Step 42340/44000 | Loss: 2.8426 | LR: 1.82e-06 | 1390.0ms/step | 14.4 samples/s
+Step 42350/44000 | Loss: 2.9020 | LR: 1.79e-06 | 1388.0ms/step | 14.4 samples/s
+Step 42360/44000 | Loss: 2.8807 | LR: 1.77e-06 | 1393.1ms/step | 14.4 samples/s
+Step 42370/44000 | Loss: 2.8975 | LR: 1.75e-06 | 1388.5ms/step | 14.4 samples/s
+Step 42380/44000 | Loss: 2.8723 | LR: 1.73e-06 | 1386.5ms/step | 14.4 samples/s
+Step 42390/44000 | Loss: 2.8710 | LR: 1.71e-06 | 1392.0ms/step | 14.4 samples/s
+Step 42400/44000 | Loss: 2.8511 | LR: 1.69e-06 | 1386.2ms/step | 14.4 samples/s
+Step 42410/44000 | Loss: 2.8969 | LR: 1.67e-06 | 1388.0ms/step | 14.4 samples/s
+Step 42420/44000 | Loss: 2.8742 | LR: 1.64e-06 | 1385.4ms/step | 14.4 samples/s
+Step 42430/44000 | Loss: 2.8785 | LR: 1.62e-06 | 1384.8ms/step | 14.4 samples/s
+Step 42440/44000 | Loss: 2.8696 | LR: 1.60e-06 | 1382.7ms/step | 14.5 samples/s
+Step 42450/44000 | Loss: 2.8940 | LR: 1.58e-06 | 1385.2ms/step | 14.4 samples/s
+Step 42460/44000 | Loss: 2.8793 | LR: 1.56e-06 | 1387.0ms/step | 14.4 samples/s
+Step 42470/44000 | Loss: 2.8719 | LR: 1.54e-06 | 1390.2ms/step | 14.4 samples/s
+Step 42480/44000 | Loss: 2.8585 | LR: 1.52e-06 | 1389.2ms/step | 14.4 samples/s
+Step 42490/44000 | Loss: 2.8795 | LR: 1.50e-06 | 1385.3ms/step | 14.4 samples/s
+Step 42500/44000 | Loss: 2.8770 | LR: 1.48e-06 | 1389.9ms/step | 14.4 samples/s
+Validation at step 42500: Loss = 2.8954, Perplexity = 18.09
+Step 42510/44000 | Loss: 2.8447 | LR: 1.46e-06 | 3592.1ms/step | 5.6 samples/s
+Step 42520/44000 | Loss: 2.8645 | LR: 1.44e-06 | 1392.2ms/step | 14.4 samples/s
+Step 42530/44000 | Loss: 2.8844 | LR: 1.42e-06 | 1386.8ms/step | 14.4 samples/s
+Step 42540/44000 | Loss: 2.8400 | LR: 1.40e-06 | 1386.1ms/step | 14.4 samples/s
+Step 42550/44000 | Loss: 2.8698 | LR: 1.39e-06 | 1387.0ms/step | 14.4 samples/s
+Step 42560/44000 | Loss: 2.8698 | LR: 1.37e-06 | 1389.7ms/step | 14.4 samples/s
+Step 42570/44000 | Loss: 2.8830 | LR: 1.35e-06 | 1384.5ms/step | 14.4 samples/s
+Step 42580/44000 | Loss: 2.8726 | LR: 1.33e-06 | 1388.8ms/step | 14.4 samples/s
+Step 42590/44000 | Loss: 2.9030 | LR: 1.31e-06 | 1387.0ms/step | 14.4 samples/s
+Step 42600/44000 | Loss: 2.8863 | LR: 1.29e-06 | 1388.9ms/step | 14.4 samples/s
+Step 42610/44000 | Loss: 2.8716 | LR: 1.27e-06 | 1390.2ms/step | 14.4 samples/s
+Step 42620/44000 | Loss: 2.8781 | LR: 1.25e-06 | 1387.0ms/step | 14.4 samples/s
+Step 42630/44000 | Loss: 2.8735 | LR: 1.24e-06 | 1389.7ms/step | 14.4 samples/s
+Step 42640/44000 | Loss: 2.8528 | LR: 1.22e-06 | 1390.2ms/step | 14.4 samples/s
+Step 42650/44000 | Loss: 2.8994 | LR: 1.20e-06 | 1389.3ms/step | 14.4 samples/s
+Step 42660/44000 | Loss: 2.8708 | LR: 1.18e-06 | 1387.0ms/step | 14.4 samples/s
+Step 42670/44000 | Loss: 2.8850 | LR: 1.17e-06 | 1387.5ms/step | 14.4 samples/s
+Step 42680/44000 | Loss: 2.9027 | LR: 1.15e-06 | 1384.3ms/step | 14.4 samples/s
+Step 42690/44000 | Loss: 2.8479 | LR: 1.13e-06 | 1387.4ms/step | 14.4 samples/s
+Step 42700/44000 | Loss: 2.8548 | LR: 1.11e-06 | 1388.8ms/step | 14.4 samples/s
+Step 42710/44000 | Loss: 2.8389 | LR: 1.10e-06 | 1383.9ms/step | 14.5 samples/s
+Step 42720/44000 | Loss: 2.8882 | LR: 1.08e-06 | 1382.8ms/step | 14.5 samples/s
+Step 42730/44000 | Loss: 2.9049 | LR: 1.06e-06 | 1384.8ms/step | 14.4 samples/s
+Step 42740/44000 | Loss: 2.8926 | LR: 1.05e-06 | 1385.1ms/step | 14.4 samples/s
+Step 42750/44000 | Loss: 2.8839 | LR: 1.03e-06 | 1395.3ms/step | 14.3 samples/s
+Validation at step 42750: Loss = 2.8954, Perplexity = 18.09
+Step 42760/44000 | Loss: 2.8667 | LR: 1.01e-06 | 3610.1ms/step | 5.5 samples/s
+Step 42770/44000 | Loss: 2.8719 | LR: 9.97e-07 | 1403.3ms/step | 14.3 samples/s
+Step 42780/44000 | Loss: 2.8468 | LR: 9.81e-07 | 1400.8ms/step | 14.3 samples/s
+Step 42790/44000 | Loss: 2.8360 | LR: 9.65e-07 | 1398.2ms/step | 14.3 samples/s
+Step 42800/44000 | Loss: 2.8706 | LR: 9.49e-07 | 1399.6ms/step | 14.3 samples/s
+Step 42810/44000 | Loss: 2.8880 | LR: 9.33e-07 | 1394.4ms/step | 14.3 samples/s
+Step 42820/44000 | Loss: 2.8811 | LR: 9.18e-07 | 1400.2ms/step | 14.3 samples/s
+Step 42830/44000 | Loss: 2.8996 | LR: 9.02e-07 | 1399.3ms/step | 14.3 samples/s
+Step 42840/44000 | Loss: 2.8647 | LR: 8.87e-07 | 1397.8ms/step | 14.3 samples/s
+Step 42850/44000 | Loss: 2.8710 | LR: 8.72e-07 | 1395.6ms/step | 14.3 samples/s
+Step 42860/44000 | Loss: 2.8771 | LR: 8.57e-07 | 1398.9ms/step | 14.3 samples/s
+Step 42870/44000 | Loss: 2.8788 | LR: 8.42e-07 | 1399.9ms/step | 14.3 samples/s
+Step 42880/44000 | Loss: 2.8910 | LR: 8.27e-07 | 1397.9ms/step | 14.3 samples/s
+Step 42890/44000 | Loss: 2.8835 | LR: 8.12e-07 | 1400.2ms/step | 14.3 samples/s
+Step 42900/44000 | Loss: 2.8589 | LR: 7.98e-07 | 1399.8ms/step | 14.3 samples/s
+Step 42910/44000 | Loss: 2.8613 | LR: 7.83e-07 | 1398.0ms/step | 14.3 samples/s

epfl_final/SparseAtt/wandb/run-20260708_045920-qlqhc3gy/files/requirements.txt

@@ -0,0 +1,141 @@
+pip==26.0.1
+nvidia-cusparselt-cu13==0.8.0
+mpmath==1.3.0
+cuda-toolkit==13.0.2
+zstandard==0.25.0
+xxhash==3.7.0
+urllib3==2.6.3
+typing_extensions==4.15.0
+triton==3.6.0
+tqdm==4.67.3
+torchinfo==1.8.0
+sympy==1.14.0
+smmap==5.0.3
+six==1.17.0
+shellingham==1.5.4
+setuptools==81.0.0
+safetensors==0.7.0
+regex==2026.4.4
+PyYAML==6.0.3
+Pygments==2.20.0
+pyarrow==24.0.0
+psutil==7.2.2
+protobuf==7.34.1
+propcache==0.4.1
+platformdirs==4.9.6
+packaging==26.2
+nvidia-nvtx==13.0.85
+nvidia-nvshmem-cu13==3.4.5
+nvidia-nvjitlink==13.0.88
+nvidia-nccl-cu13==2.28.9
+nvidia-curand==10.4.0.35
+nvidia-cufile==1.15.1.6
+nvidia-cuda-runtime==13.0.96
+nvidia-cuda-nvrtc==13.0.88
+nvidia-cuda-cupti==13.0.85
+nvidia-cublas==13.1.0.3
+numpy==2.4.4
+networkx==3.6.1
+multidict==6.7.1
+mdurl==0.1.2
+MarkupSafe==3.0.3
+idna==3.13
+hf-xet==1.4.3
+h11==0.16.0
+fsspec==2026.2.0
+frozenlist==1.8.0
+filelock==3.29.0
+dill==0.4.1
+cuda-pathfinder==1.5.4
+click==8.3.3
+charset-normalizer==3.4.7
+certifi==2026.4.22
+attrs==26.1.0
+annotated-types==0.7.0
+annotated-doc==0.0.4
+aiohappyeyeballs==2.6.1
+yarl==1.23.0
+typing-inspection==0.4.2
+sentry-sdk==2.58.0
+requests==2.33.1
+python-dateutil==2.9.0.post0
+pydantic_core==2.46.3
+nvidia-cusparse==12.6.3.3
+nvidia-cufft==12.0.0.61
+nvidia-cudnn-cu13==9.19.0.56
+multiprocess==0.70.19
+markdown-it-py==4.0.0
+Jinja2==3.1.6
+httpcore==1.0.9
+gitdb==4.0.12
+cuda-bindings==13.2.0
+anyio==4.13.0
+aiosignal==1.4.0
+rich==15.0.0
+pydantic==2.13.3
+pandas==3.0.2
+nvidia-cusolver==12.0.4.66
+httpx==0.28.1
+GitPython==3.1.49
+aiohttp==3.13.5
+wandb==0.26.1
+typer==0.25.1
+torch==2.11.0
+huggingface_hub==1.13.0
+tokenizers==0.22.2
+datasets==4.8.5
+accelerate==1.13.0
+transformers==5.7.0
+word2number==1.1
+sqlitedict==2.1.0
+pytz==2026.2
+threadpoolctl==3.6.0
+tcolorpy==0.1.7
+tabulate==0.10.0
+scipy==1.17.1
+portalocker==3.2.0
+pathvalidate==3.3.1
+narwhals==2.22.1
+more-itertools==11.1.0
+lxml==6.1.1
+joblib==1.5.3
+colorama==0.4.6
+chardet==6.0.0.post1
+absl-py==2.4.0
+scikit-learn==1.9.0
+sacrebleu==2.6.0
+nltk==3.9.4
+mbstrdecoder==1.1.5
+typepy==1.3.5
+rouge_score==0.1.2
+DataProperty==1.1.1
+tabledata==1.3.5
+evaluate==0.4.6
+pytablewriter==1.2.1
+lm_eval==0.4.12
+pluggy==1.6.0
+iniconfig==2.3.0
+pytest==9.1.0
+py-spy==0.4.2
+pillow==12.2.0
+tifffile==2026.6.1
+imagecodecs==2026.6.26
+zipp==4.1.0
+importlib_metadata==9.0.0
+derivative==0.6.3
+pysindy==2.1.0
+pyparsing==3.3.2
+kiwisolver==1.5.0
+fonttools==4.63.0
+cycler==0.12.1
+contourpy==1.3.3
+matplotlib==3.11.0
+sortedcontainers==2.4.0
+wcwidth==0.8.2
+sigfig==1.3.19
+importlib_resources==7.1.0
+enum-tools==0.13.0
+cramjam==2.11.0
+python-snappy==0.7.3
+compact-json==1.8.2
+numbers-parser==4.18.5

epfl_final/SparseAtt/wandb/run-20260708_045920-qlqhc3gy/files/wandb-metadata.json

@@ -0,0 +1,44 @@
+{
+  "os":  "Linux-6.8.0-124-generic-x86_64-with-glibc2.35",
+  "python":  "CPython 3.13.13",
+  "startedAt":  "2026-07-08T04:59:20.834106Z",
+  "args":  [
+    "--config",
+    "configs/ladder8k/dense.json"
+  ],
+  "program":  "/mloscratch/homes/navasard/sparse/SparseAtt/train.py",
+  "codePath":  "train.py",
+  "codePathLocal":  "train.py",
+  "git":  {
+    "remote":  "https://github.com/rubentium/TokenPrune.git",
+    "commit":  "88298c7a418622530ca5426b22b1e5ebd10ca0cb"
+  },
+  "email":  "ruben.navasardyan@mail.utoronto.ca",
+  "root":  "/mloscratch/homes/navasard/sparse/SparseAtt",
+  "host":  "h100-workspace-1-gpu-0-0",
+  "executable":  "/mloscratch/homes/navasard/conda/bin/python",
+  "cpu_count":  96,
+  "cpu_count_logical":  192,
+  "gpu":  "NVIDIA H100 80GB HBM3",
+  "gpu_count":  1,
+  "disk":  {
+    "/":  {
+      "total":  "9374931091456",
+      "used":  "3858025717760"
+    }
+  },
+  "memory":  {
+    "total":  "1623005732864"
+  },
+  "gpu_nvidia":  [
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-3f5bace4-0066-71eb-863d-73a3a6d79e3b"
+    }
+  ],
+  "cudaVersion":  "13.0",
+  "writerId":  "0w7lkft2stdrzhjtwih8k3en8w1vn31a"
+}

epfl_final/SparseAtt/wandb/run-20260708_045920-qlqhc3gy/logs/debug-internal.log

@@ -0,0 +1,513 @@
+{"time":"2026-07-08T04:59:21.434170127Z","level":"INFO","msg":"wandb-core"}
+{"time":"2026-07-08T04:59:21.43947287Z","level":"INFO","msg":"stream: starting","core version":"0.26.1"}
+{"time":"2026-07-08T04:59:21.857105678Z","level":"INFO","msg":"stream: created new stream","id":"qlqhc3gy"}
+{"time":"2026-07-08T04:59:21.857182222Z","level":"INFO","msg":"handler: started"}
+{"time":"2026-07-08T04:59:21.859152166Z","level":"INFO","msg":"stream: started"}
+{"time":"2026-07-08T04:59:21.860059792Z","level":"INFO","msg":"sender: started"}
+{"time":"2026-07-08T04:59:21.860063758Z","level":"INFO","msg":"writer: started","stream_id":"qlqhc3gy"}
+{"time":"2026-07-08T04:59:22.541639888Z","level":"INFO","msg":"filestream: sending request","total_files":1}
+{"time":"2026-07-08T04:59:22.855283189Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T04:59:37.541966158Z","level":"INFO","msg":"filestream: sending request","total_files":1,"console_offset":0,"console_lines":1,"uploaded_len":2}
+{"time":"2026-07-08T04:59:37.815936141Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T04:59:52.542432627Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":0,"events_lines":2}
+{"time":"2026-07-08T04:59:52.8254775Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:00:07.542154674Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":2,"events_lines":2}
+{"time":"2026-07-08T05:00:07.778811146Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:00:07.979500937Z","level":"WARN","msg":"handler: ignoring partial history record","step":41000,"current":41321}
+{"time":"2026-07-08T05:00:22.542081107Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":4,"events_lines":2,"console_offset":1,"console_lines":2}
+{"time":"2026-07-08T05:00:22.797093285Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:00:37.541766059Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":6,"events_lines":2}
+{"time":"2026-07-08T05:00:37.771603772Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:00:52.541781505Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":8,"events_lines":2}
+{"time":"2026-07-08T05:00:52.831175514Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:01:07.542029908Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":10,"events_lines":2}
+{"time":"2026-07-08T05:01:07.895718293Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:01:22.542576257Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":12,"events_lines":2}
+{"time":"2026-07-08T05:01:22.873902109Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:01:37.541621513Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":14,"events_lines":2}
+{"time":"2026-07-08T05:01:37.815067634Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:01:52.54225347Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":16,"events_lines":2}
+{"time":"2026-07-08T05:01:52.844195708Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:02:07.542328645Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":18,"events_lines":2}
+{"time":"2026-07-08T05:02:07.750618342Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:02:22.542328848Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":20,"events_lines":2}
+{"time":"2026-07-08T05:02:22.791304856Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:02:37.542520877Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":22,"events_lines":2}
+{"time":"2026-07-08T05:02:37.749197914Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:02:52.54213297Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":24,"events_lines":2}
+{"time":"2026-07-08T05:02:52.879610273Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:03:07.541778533Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":26,"events_lines":2}
+{"time":"2026-07-08T05:03:07.855886659Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:03:22.54260471Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":28,"events_lines":2}
+{"time":"2026-07-08T05:03:22.820462552Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:03:37.5421349Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":30,"events_lines":2}
+{"time":"2026-07-08T05:03:37.783851242Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:03:52.5421127Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":32,"events_lines":2}
+{"time":"2026-07-08T05:03:52.827008258Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:04:07.542587312Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":34,"events_lines":2}
+{"time":"2026-07-08T05:04:07.841027107Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:04:22.541730354Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":36,"events_lines":2}
+{"time":"2026-07-08T05:04:22.947190764Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:04:37.542037485Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":38,"events_lines":2}
+{"time":"2026-07-08T05:04:37.8281623Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:04:52.5425805Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":40,"events_lines":2}
+{"time":"2026-07-08T05:04:52.749909087Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:05:07.542502103Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":42,"events_lines":2}
+{"time":"2026-07-08T05:05:07.82968109Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:05:22.542080189Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":44,"events_lines":2}
+{"time":"2026-07-08T05:05:22.773073107Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:05:37.541812637Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":46,"events_lines":2}
+{"time":"2026-07-08T05:05:37.772391674Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:05:52.542161175Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":48,"events_lines":2}
+{"time":"2026-07-08T05:05:52.773703243Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:06:07.542705386Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":50,"events_lines":2}
+{"time":"2026-07-08T05:06:07.814732375Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:06:22.542592554Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":52,"events_lines":2}
+{"time":"2026-07-08T05:06:22.758034195Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:06:37.542187295Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":54,"events_lines":2}
+{"time":"2026-07-08T05:06:37.841225307Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:06:52.542219579Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":56,"events_lines":2}
+{"time":"2026-07-08T05:06:52.873045927Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:07:07.541902011Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":58,"events_lines":2}
+{"time":"2026-07-08T05:07:07.743187864Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:07:22.542373884Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":60,"events_lines":2}
+{"time":"2026-07-08T05:07:22.797271929Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:07:37.542397857Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":62,"events_lines":2}
+{"time":"2026-07-08T05:07:37.778583419Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:07:52.541945698Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":64,"events_lines":2}
+{"time":"2026-07-08T05:07:52.787838646Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:08:07.541702451Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":66,"events_lines":2}
+{"time":"2026-07-08T05:08:07.768150492Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:08:22.541708418Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":68,"events_lines":2}
+{"time":"2026-07-08T05:08:22.880761261Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:08:37.542476861Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":70,"events_lines":2}
+{"time":"2026-07-08T05:08:37.775447041Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:08:52.542724917Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":72,"events_lines":2}
+{"time":"2026-07-08T05:08:52.80775773Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:09:07.541878688Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":74,"events_lines":2}
+{"time":"2026-07-08T05:09:07.784811542Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:09:22.541821667Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":76,"events_lines":2}
+{"time":"2026-07-08T05:09:22.860429184Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:09:37.542051165Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":78,"events_lines":2}
+{"time":"2026-07-08T05:09:37.872091283Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:09:52.542751727Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":80,"events_lines":2}
+{"time":"2026-07-08T05:09:52.825520429Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:10:07.542547187Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":82,"events_lines":2}
+{"time":"2026-07-08T05:10:07.777025109Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:10:22.542395565Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":84,"events_lines":2}
+{"time":"2026-07-08T05:10:22.79456481Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:10:37.542042132Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":86,"events_lines":2,"console_offset":3,"console_lines":1}
+{"time":"2026-07-08T05:10:37.824268664Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:10:52.542596406Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":88,"events_lines":2}
+{"time":"2026-07-08T05:10:52.818222135Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:11:07.541794292Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":90,"events_lines":2}
+{"time":"2026-07-08T05:11:07.751857628Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:11:22.54189668Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":92,"events_lines":2}
+{"time":"2026-07-08T05:11:22.753146078Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:11:29.875264961Z","level":"WARN","msg":"handler: ignoring partial history record","step":41010,"current":41321}
+{"time":"2026-07-08T05:11:37.542032248Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":94,"events_lines":2,"console_offset":4,"console_lines":2}
+{"time":"2026-07-08T05:11:37.770990463Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:11:43.65662946Z","level":"WARN","msg":"handler: ignoring partial history record","step":41020,"current":41321}
+{"time":"2026-07-08T05:11:52.542701423Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":96,"events_lines":2,"console_offset":6,"console_lines":1}
+{"time":"2026-07-08T05:11:52.748133589Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:11:57.44399848Z","level":"WARN","msg":"handler: ignoring partial history record","step":41030,"current":41321}
+{"time":"2026-07-08T05:12:07.54192673Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":98,"events_lines":2,"console_offset":7,"console_lines":3}
+{"time":"2026-07-08T05:12:07.851474751Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:12:11.270903499Z","level":"WARN","msg":"handler: ignoring partial history record","step":41040,"current":41321}
+{"time":"2026-07-08T05:12:22.541705925Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":100,"events_lines":2,"console_offset":10,"console_lines":2}
+{"time":"2026-07-08T05:12:22.763110686Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:12:25.114697411Z","level":"WARN","msg":"handler: ignoring partial history record","step":41050,"current":41321}
+{"time":"2026-07-08T05:12:37.542468069Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":102,"events_lines":2,"console_offset":12,"console_lines":2}
+{"time":"2026-07-08T05:12:37.824144128Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:12:38.943246346Z","level":"WARN","msg":"handler: ignoring partial history record","step":41060,"current":41321}
+{"time":"2026-07-08T05:12:52.541713617Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":104,"events_lines":2,"console_offset":14,"console_lines":2}
+{"time":"2026-07-08T05:12:52.74469379Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:12:52.814744459Z","level":"WARN","msg":"handler: ignoring partial history record","step":41070,"current":41321}
+{"time":"2026-07-08T05:13:06.657327933Z","level":"WARN","msg":"handler: ignoring partial history record","step":41080,"current":41321}
+{"time":"2026-07-08T05:13:07.541943015Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":106,"events_lines":2,"console_offset":16,"console_lines":3}
+{"time":"2026-07-08T05:13:07.750533109Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:13:20.508777162Z","level":"WARN","msg":"handler: ignoring partial history record","step":41090,"current":41321}
+{"time":"2026-07-08T05:13:22.542180164Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":108,"events_lines":2,"console_offset":19,"console_lines":2}
+{"time":"2026-07-08T05:13:22.762332804Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:13:34.355958085Z","level":"WARN","msg":"handler: ignoring partial history record","step":41100,"current":41321}
+{"time":"2026-07-08T05:13:37.542259958Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":110,"events_lines":2,"console_offset":21,"console_lines":2}
+{"time":"2026-07-08T05:13:37.810431801Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:13:48.235895823Z","level":"WARN","msg":"handler: ignoring partial history record","step":41110,"current":41321}
+{"time":"2026-07-08T05:13:52.542341197Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":112,"events_lines":2,"console_offset":23,"console_lines":2}
+{"time":"2026-07-08T05:13:52.741260885Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:14:02.11592558Z","level":"WARN","msg":"handler: ignoring partial history record","step":41120,"current":41321}
+{"time":"2026-07-08T05:14:07.542430261Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":114,"events_lines":2,"console_offset":25,"console_lines":3}
+{"time":"2026-07-08T05:14:07.838890151Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:14:15.978307189Z","level":"WARN","msg":"handler: ignoring partial history record","step":41130,"current":41321}
+{"time":"2026-07-08T05:14:22.542694437Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":116,"events_lines":2,"console_offset":28,"console_lines":1}
+{"time":"2026-07-08T05:14:22.751198241Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:14:29.832891424Z","level":"WARN","msg":"handler: ignoring partial history record","step":41140,"current":41321}
+{"time":"2026-07-08T05:14:37.54257622Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":118,"events_lines":2,"console_offset":29,"console_lines":3}
+{"time":"2026-07-08T05:14:37.829149822Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:14:43.717070648Z","level":"WARN","msg":"handler: ignoring partial history record","step":41150,"current":41321}
+{"time":"2026-07-08T05:14:52.542505395Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":120,"events_lines":2,"console_offset":32,"console_lines":1}
+{"time":"2026-07-08T05:14:52.801990608Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:14:57.596293499Z","level":"WARN","msg":"handler: ignoring partial history record","step":41160,"current":41321}
+{"time":"2026-07-08T05:15:07.541757797Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":122,"events_lines":2,"console_offset":33,"console_lines":3}
+{"time":"2026-07-08T05:15:07.883113293Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:15:11.464186231Z","level":"WARN","msg":"handler: ignoring partial history record","step":41170,"current":41321}
+{"time":"2026-07-08T05:15:22.542372537Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":124,"events_lines":2,"console_offset":36,"console_lines":2}
+{"time":"2026-07-08T05:15:22.740551327Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:15:25.338638235Z","level":"WARN","msg":"handler: ignoring partial history record","step":41180,"current":41321}
+{"time":"2026-07-08T05:15:37.541902024Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":126,"events_lines":2,"console_offset":38,"console_lines":2}
+{"time":"2026-07-08T05:15:37.8331382Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:15:39.214451626Z","level":"WARN","msg":"handler: ignoring partial history record","step":41190,"current":41321}
+{"time":"2026-07-08T05:15:52.541794593Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":128,"events_lines":2,"console_offset":40,"console_lines":2}
+{"time":"2026-07-08T05:15:52.805565347Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:15:53.090576121Z","level":"WARN","msg":"handler: ignoring partial history record","step":41200,"current":41321}
+{"time":"2026-07-08T05:16:06.974714094Z","level":"WARN","msg":"handler: ignoring partial history record","step":41210,"current":41321}
+{"time":"2026-07-08T05:16:07.542000452Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":130,"events_lines":2,"console_offset":42,"console_lines":3}
+{"time":"2026-07-08T05:16:07.798835992Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:16:20.856277275Z","level":"WARN","msg":"handler: ignoring partial history record","step":41220,"current":41321}
+{"time":"2026-07-08T05:16:22.542626809Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":132,"events_lines":2,"console_offset":45,"console_lines":2}
+{"time":"2026-07-08T05:16:22.78452186Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:16:34.711591332Z","level":"WARN","msg":"handler: ignoring partial history record","step":41230,"current":41321}
+{"time":"2026-07-08T05:16:37.542241287Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":134,"events_lines":2,"console_offset":47,"console_lines":2}
+{"time":"2026-07-08T05:16:37.816748154Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:16:48.574769674Z","level":"WARN","msg":"handler: ignoring partial history record","step":41240,"current":41321}
+{"time":"2026-07-08T05:16:52.541831127Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":136,"events_lines":2,"console_offset":49,"console_lines":2}
+{"time":"2026-07-08T05:16:52.814101312Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:17:02.439027226Z","level":"WARN","msg":"handler: ignoring partial history record","step":41250,"current":41321}
+{"time":"2026-07-08T05:17:07.542263485Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":138,"events_lines":2,"console_offset":51,"console_lines":3}
+{"time":"2026-07-08T05:17:07.869015343Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:17:22.542374934Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":140,"events_lines":2}
+{"time":"2026-07-08T05:17:22.768587303Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:17:24.483715117Z","level":"WARN","msg":"handler: ignoring partial history record","step":41250,"current":41321}
+{"time":"2026-07-08T05:17:37.542101287Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":142,"events_lines":2,"console_offset":54,"console_lines":1}
+{"time":"2026-07-08T05:17:37.799011323Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:17:38.437871639Z","level":"WARN","msg":"handler: ignoring partial history record","step":41260,"current":41321}
+{"time":"2026-07-08T05:17:52.396683077Z","level":"WARN","msg":"handler: ignoring partial history record","step":41270,"current":41321}
+{"time":"2026-07-08T05:17:52.542053048Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":144,"events_lines":2,"console_offset":55,"console_lines":3}
+{"time":"2026-07-08T05:17:52.730497857Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:18:06.37018934Z","level":"WARN","msg":"handler: ignoring partial history record","step":41280,"current":41321}
+{"time":"2026-07-08T05:18:07.541695175Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":146,"events_lines":2,"console_offset":58,"console_lines":2}
+{"time":"2026-07-08T05:18:07.863764487Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:18:20.294188944Z","level":"WARN","msg":"handler: ignoring partial history record","step":41290,"current":41321}
+{"time":"2026-07-08T05:18:22.541878481Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":148,"events_lines":2,"console_offset":60,"console_lines":2}
+{"time":"2026-07-08T05:18:22.763616455Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:18:34.23494205Z","level":"WARN","msg":"handler: ignoring partial history record","step":41300,"current":41321}
+{"time":"2026-07-08T05:18:37.542410128Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":150,"events_lines":2,"console_offset":62,"console_lines":2}
+{"time":"2026-07-08T05:18:37.790897839Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:18:48.18120614Z","level":"WARN","msg":"handler: ignoring partial history record","step":41310,"current":41321}
+{"time":"2026-07-08T05:18:52.542274758Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":152,"events_lines":2,"console_offset":64,"console_lines":2}
+{"time":"2026-07-08T05:18:52.768218927Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:19:02.1363651Z","level":"WARN","msg":"handler: ignoring partial history record","step":41320,"current":41321}
+{"time":"2026-07-08T05:19:07.542346918Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":154,"events_lines":2,"console_offset":66,"console_lines":3}
+{"time":"2026-07-08T05:19:07.799054251Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:19:22.542287852Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":156,"events_lines":2,"console_offset":69,"console_lines":1}
+{"time":"2026-07-08T05:19:22.767525866Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:19:37.542021346Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4132,"history_lines":1,"events_offset":158,"events_lines":2,"console_offset":70,"console_lines":1}
+{"time":"2026-07-08T05:19:37.796824554Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:19:52.542369849Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4133,"history_lines":1,"events_offset":160,"events_lines":2,"console_offset":71,"console_lines":1}
+{"time":"2026-07-08T05:19:52.942114698Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:20:07.541916795Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4134,"history_lines":1,"events_offset":162,"events_lines":2,"console_offset":72,"console_lines":1}
+{"time":"2026-07-08T05:20:07.937392094Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:20:22.542678675Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4135,"history_lines":1,"events_offset":164,"events_lines":2,"console_offset":73,"console_lines":1}
+{"time":"2026-07-08T05:20:22.84995229Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:20:37.541867586Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4136,"history_lines":1,"events_offset":166,"events_lines":2,"console_offset":74,"console_lines":1}
+{"time":"2026-07-08T05:20:37.836794133Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:20:52.54186207Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4137,"history_lines":1,"events_offset":168,"events_lines":2,"console_offset":75,"console_lines":1}
+{"time":"2026-07-08T05:20:52.818044681Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:21:07.54198083Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4138,"history_lines":1,"events_offset":170,"events_lines":2,"console_offset":76,"console_lines":1}
+{"time":"2026-07-08T05:21:07.855986033Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:21:22.542374981Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4139,"history_lines":2,"events_offset":172,"events_lines":2,"console_offset":77,"console_lines":2}
+{"time":"2026-07-08T05:21:22.825032557Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:21:37.542269145Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4141,"history_lines":1,"events_offset":174,"events_lines":2,"console_offset":79,"console_lines":1}
+{"time":"2026-07-08T05:21:37.851602617Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:21:52.541971102Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4142,"history_lines":1,"events_offset":176,"events_lines":2,"console_offset":80,"console_lines":1}
+{"time":"2026-07-08T05:21:52.776818744Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:22:07.542291973Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4143,"history_lines":1,"events_offset":178,"events_lines":2,"console_offset":81,"console_lines":1}
+{"time":"2026-07-08T05:22:07.821130503Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:22:22.542158463Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4144,"history_lines":1,"events_offset":180,"events_lines":2,"console_offset":82,"console_lines":1}
+{"time":"2026-07-08T05:22:23.059264849Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:22:37.542674373Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4145,"history_lines":1,"events_offset":182,"events_lines":2,"console_offset":83,"console_lines":1}
+{"time":"2026-07-08T05:22:37.78535565Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:22:52.542259202Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4146,"history_lines":1,"events_offset":184,"events_lines":2,"console_offset":84,"console_lines":1}
+{"time":"2026-07-08T05:22:52.875384061Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:23:07.542779229Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4147,"history_lines":1,"events_offset":186,"events_lines":2,"console_offset":85,"console_lines":1}
+{"time":"2026-07-08T05:23:07.785999451Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:23:22.542746199Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4148,"history_lines":1,"events_offset":188,"events_lines":2,"console_offset":86,"console_lines":1}
+{"time":"2026-07-08T05:23:22.844981522Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:23:37.542694051Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":190,"events_lines":2,"console_offset":87,"console_lines":1}
+{"time":"2026-07-08T05:23:37.858288099Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:23:52.541980944Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4149,"history_lines":1,"events_offset":192,"events_lines":2,"console_offset":88,"console_lines":1}
+{"time":"2026-07-08T05:23:52.85103616Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:24:07.542192841Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4150,"history_lines":1,"events_offset":194,"events_lines":2,"console_offset":89,"console_lines":1}
+{"time":"2026-07-08T05:24:07.79919423Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:24:22.542437376Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4151,"history_lines":1,"events_offset":196,"events_lines":2,"console_offset":90,"console_lines":1}
+{"time":"2026-07-08T05:24:22.930092812Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:24:37.541839742Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4152,"history_lines":1,"events_offset":198,"events_lines":2,"console_offset":91,"console_lines":1}
+{"time":"2026-07-08T05:24:38.025128617Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:24:52.542567639Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4153,"history_lines":1,"events_offset":200,"events_lines":2,"console_offset":92,"console_lines":1}
+{"time":"2026-07-08T05:24:52.783931958Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:25:07.541996095Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4154,"history_lines":1,"events_offset":202,"events_lines":2,"console_offset":93,"console_lines":1}
+{"time":"2026-07-08T05:25:07.822447581Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:25:22.542077206Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4155,"history_lines":1,"events_offset":204,"events_lines":2,"console_offset":94,"console_lines":1}
+{"time":"2026-07-08T05:25:22.89543301Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:25:37.541935314Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4156,"history_lines":1,"events_offset":206,"events_lines":2,"console_offset":95,"console_lines":1}
+{"time":"2026-07-08T05:25:37.797159066Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:25:52.542798823Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4157,"history_lines":1,"events_offset":208,"events_lines":2,"console_offset":96,"console_lines":1}
+{"time":"2026-07-08T05:25:53.359342922Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:26:07.542742769Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4158,"history_lines":1,"events_offset":210,"events_lines":2,"console_offset":97,"console_lines":1}
+{"time":"2026-07-08T05:26:07.89600215Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:26:22.542408109Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4159,"history_lines":1,"events_offset":212,"events_lines":2,"console_offset":98,"console_lines":1}
+{"time":"2026-07-08T05:26:22.873013091Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:26:37.542209341Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4160,"history_lines":1,"events_offset":214,"events_lines":2,"console_offset":99,"console_lines":1}
+{"time":"2026-07-08T05:26:37.90689608Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:26:52.54244297Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4161,"history_lines":2,"events_offset":216,"events_lines":2,"console_offset":100,"console_lines":2}
+{"time":"2026-07-08T05:26:52.87162145Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:27:07.542702858Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4163,"history_lines":1,"events_offset":218,"events_lines":2,"console_offset":102,"console_lines":1}
+{"time":"2026-07-08T05:27:07.843959207Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:27:22.542498822Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4164,"history_lines":1,"events_offset":220,"events_lines":2,"console_offset":103,"console_lines":1}
+{"time":"2026-07-08T05:27:22.900391137Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:27:37.542158623Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4165,"history_lines":1,"events_offset":222,"events_lines":2,"console_offset":104,"console_lines":1}
+{"time":"2026-07-08T05:27:37.830597707Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:27:52.542099004Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4166,"history_lines":1,"events_offset":224,"events_lines":2,"console_offset":105,"console_lines":1}
+{"time":"2026-07-08T05:27:52.93932886Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:28:07.542358551Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4167,"history_lines":1,"events_offset":226,"events_lines":2,"console_offset":106,"console_lines":1}
+{"time":"2026-07-08T05:28:07.993873729Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:28:22.542030109Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4168,"history_lines":1,"events_offset":228,"events_lines":2,"console_offset":107,"console_lines":1}
+{"time":"2026-07-08T05:28:22.972968534Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:28:37.542745248Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4169,"history_lines":1,"events_offset":230,"events_lines":2,"console_offset":108,"console_lines":1}
+{"time":"2026-07-08T05:28:37.83911188Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:28:52.542356756Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4170,"history_lines":1,"events_offset":232,"events_lines":2,"console_offset":109,"console_lines":1}
+{"time":"2026-07-08T05:28:52.758482325Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:29:07.542052702Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4171,"history_lines":1,"events_offset":234,"events_lines":2,"console_offset":110,"console_lines":1}
+{"time":"2026-07-08T05:29:07.927235924Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:29:22.542318699Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4172,"history_lines":1,"events_offset":236,"events_lines":2,"console_offset":111,"console_lines":1}
+{"time":"2026-07-08T05:29:22.820627633Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:29:37.542257206Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4173,"history_lines":1,"events_offset":238,"events_lines":2,"console_offset":112,"console_lines":1}
+{"time":"2026-07-08T05:29:37.859584835Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:29:52.542758563Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":240,"events_lines":2,"console_offset":113,"console_lines":1}
+{"time":"2026-07-08T05:29:52.85608802Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:30:07.541938536Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4174,"history_lines":1,"events_offset":242,"events_lines":2,"console_offset":114,"console_lines":1}
+{"time":"2026-07-08T05:30:08.012108119Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:30:22.541846887Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4175,"history_lines":1,"events_offset":244,"events_lines":2,"console_offset":115,"console_lines":1}
+{"time":"2026-07-08T05:30:22.860140673Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:30:37.542261796Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4176,"history_lines":1,"events_offset":246,"events_lines":2,"console_offset":116,"console_lines":1}
+{"time":"2026-07-08T05:30:37.906916038Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:30:52.541976809Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4177,"history_lines":1,"events_offset":248,"events_lines":2,"console_offset":117,"console_lines":1}
+{"time":"2026-07-08T05:30:52.836661726Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:31:07.542709885Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4178,"history_lines":1,"events_offset":250,"events_lines":2,"console_offset":118,"console_lines":1}
+{"time":"2026-07-08T05:31:08.078734065Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:31:22.542090761Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4179,"history_lines":1,"events_offset":252,"events_lines":2,"console_offset":119,"console_lines":1}
+{"time":"2026-07-08T05:31:22.974536539Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:31:37.542050692Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4180,"history_lines":1,"events_offset":254,"events_lines":2,"console_offset":120,"console_lines":1}
+{"time":"2026-07-08T05:31:37.892900088Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:31:52.542541357Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4181,"history_lines":1,"events_offset":256,"events_lines":2,"console_offset":121,"console_lines":1}
+{"time":"2026-07-08T05:31:52.881748564Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:32:07.542580055Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4182,"history_lines":1,"events_offset":258,"events_lines":2,"console_offset":122,"console_lines":1}
+{"time":"2026-07-08T05:32:08.051932717Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:32:22.54256226Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4183,"history_lines":2,"events_offset":260,"events_lines":2,"console_offset":123,"console_lines":2}
+{"time":"2026-07-08T05:32:22.908948255Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:32:37.541829796Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4185,"history_lines":1,"events_offset":262,"events_lines":2,"console_offset":125,"console_lines":1}
+{"time":"2026-07-08T05:32:37.878710808Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:32:52.542524055Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4186,"history_lines":1,"events_offset":264,"events_lines":2,"console_offset":126,"console_lines":1}
+{"time":"2026-07-08T05:32:52.826786341Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:33:07.542564156Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4187,"history_lines":1,"events_offset":266,"events_lines":2,"console_offset":127,"console_lines":1}
+{"time":"2026-07-08T05:33:07.88229708Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:33:22.542717997Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4188,"history_lines":1,"events_offset":268,"events_lines":2,"console_offset":128,"console_lines":1}
+{"time":"2026-07-08T05:33:22.840614999Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:33:37.542413644Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4189,"history_lines":1,"events_offset":270,"events_lines":2,"console_offset":129,"console_lines":1}
+{"time":"2026-07-08T05:33:37.89288846Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:33:52.542253746Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4190,"history_lines":1,"events_offset":272,"events_lines":2,"console_offset":130,"console_lines":1}
+{"time":"2026-07-08T05:33:52.919638127Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:34:07.542044084Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4191,"history_lines":1,"events_offset":274,"events_lines":2,"console_offset":131,"console_lines":1}
+{"time":"2026-07-08T05:34:07.853132329Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:34:22.542179858Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4192,"history_lines":1,"events_offset":276,"events_lines":2,"console_offset":132,"console_lines":1}
+{"time":"2026-07-08T05:34:22.989986261Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:34:37.542383593Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4193,"history_lines":1,"events_offset":278,"events_lines":2,"console_offset":133,"console_lines":1}
+{"time":"2026-07-08T05:34:37.765753679Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:34:52.541978679Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4194,"history_lines":1,"events_offset":280,"events_lines":2,"console_offset":134,"console_lines":1}
+{"time":"2026-07-08T05:34:52.806061518Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:35:07.542206741Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4195,"history_lines":1,"events_offset":282,"events_lines":2,"console_offset":135,"console_lines":1}
+{"time":"2026-07-08T05:35:07.855857917Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:35:22.542517177Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4196,"history_lines":1,"events_offset":284,"events_lines":2,"console_offset":136,"console_lines":1}
+{"time":"2026-07-08T05:35:22.886403157Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:35:37.541989289Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4197,"history_lines":2,"events_offset":286,"events_lines":2,"console_offset":137,"console_lines":2}
+{"time":"2026-07-08T05:35:37.847726157Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:35:52.541960517Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":288,"events_lines":2}
+{"time":"2026-07-08T05:35:53.279886449Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:36:07.542275699Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":290,"events_lines":2,"console_offset":139,"console_lines":2}
+{"time":"2026-07-08T05:36:07.841876116Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:36:22.541869423Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4199,"history_lines":1,"events_offset":292,"events_lines":2,"console_offset":141,"console_lines":1}
+{"time":"2026-07-08T05:36:22.946250009Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:36:37.54281681Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4200,"history_lines":1,"events_offset":294,"events_lines":2,"console_offset":142,"console_lines":1}
+{"time":"2026-07-08T05:36:38.051945036Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:36:52.54191253Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4201,"history_lines":1,"events_offset":296,"events_lines":2,"console_offset":143,"console_lines":1}
+{"time":"2026-07-08T05:36:52.820747304Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:37:07.541799331Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4202,"history_lines":1,"events_offset":298,"events_lines":2,"console_offset":144,"console_lines":1}
+{"time":"2026-07-08T05:37:07.913347441Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:37:22.542760879Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4203,"history_lines":1,"events_offset":300,"events_lines":2,"console_offset":145,"console_lines":1}
+{"time":"2026-07-08T05:37:22.813043673Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:37:37.542816353Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4204,"history_lines":1,"events_offset":302,"events_lines":2,"console_offset":146,"console_lines":1}
+{"time":"2026-07-08T05:37:37.813904351Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:37:52.542716445Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4205,"history_lines":1,"events_offset":304,"events_lines":2,"console_offset":147,"console_lines":1}
+{"time":"2026-07-08T05:37:52.961277023Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:38:07.542086295Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4206,"history_lines":1,"events_offset":306,"events_lines":2,"console_offset":148,"console_lines":1}
+{"time":"2026-07-08T05:38:07.872117539Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:38:22.542041239Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4207,"history_lines":2,"events_offset":308,"events_lines":2,"console_offset":149,"console_lines":2}
+{"time":"2026-07-08T05:38:22.953330439Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:38:37.542513425Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4209,"history_lines":1,"events_offset":310,"events_lines":2,"console_offset":151,"console_lines":1}
+{"time":"2026-07-08T05:38:37.891422422Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:38:52.541979659Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4210,"history_lines":1,"events_offset":312,"events_lines":2,"console_offset":152,"console_lines":1}
+{"time":"2026-07-08T05:38:52.809622417Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:39:07.542562561Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4211,"history_lines":1,"events_offset":314,"events_lines":2,"console_offset":153,"console_lines":1}
+{"time":"2026-07-08T05:39:07.865747234Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:39:22.542585246Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4212,"history_lines":1,"events_offset":316,"events_lines":2,"console_offset":154,"console_lines":1}
+{"time":"2026-07-08T05:39:22.881392567Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:39:37.542482263Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4213,"history_lines":1,"events_offset":318,"events_lines":2,"console_offset":155,"console_lines":1}
+{"time":"2026-07-08T05:39:37.898535517Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:39:52.542249022Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4214,"history_lines":1,"events_offset":320,"events_lines":2,"console_offset":156,"console_lines":1}
+{"time":"2026-07-08T05:39:52.919268337Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:40:07.541853487Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4215,"history_lines":1,"events_offset":322,"events_lines":2,"console_offset":157,"console_lines":1}
+{"time":"2026-07-08T05:40:07.939609952Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:40:22.541875537Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4216,"history_lines":1,"events_offset":324,"events_lines":2,"console_offset":158,"console_lines":1}
+{"time":"2026-07-08T05:40:22.933291254Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:40:37.542195272Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4217,"history_lines":1,"events_offset":326,"events_lines":2,"console_offset":159,"console_lines":1}
+{"time":"2026-07-08T05:40:37.886342867Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:40:52.542381748Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4218,"history_lines":1,"events_offset":328,"events_lines":2,"console_offset":160,"console_lines":1}
+{"time":"2026-07-08T05:40:53.480367261Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:41:07.542554062Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4219,"history_lines":1,"events_offset":330,"events_lines":2,"console_offset":161,"console_lines":1}
+{"time":"2026-07-08T05:41:08.122440293Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:41:22.54193251Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4220,"history_lines":2,"events_offset":332,"events_lines":2,"console_offset":162,"console_lines":2}
+{"time":"2026-07-08T05:41:22.986507585Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:41:37.542411003Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4222,"history_lines":1,"events_offset":334,"events_lines":2,"console_offset":164,"console_lines":1}
+{"time":"2026-07-08T05:41:37.911413814Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:41:52.542412462Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4223,"history_lines":1,"events_offset":336,"events_lines":2,"console_offset":165,"console_lines":1}
+{"time":"2026-07-08T05:41:52.932975484Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:42:07.542073221Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":338,"events_lines":2}
+{"time":"2026-07-08T05:42:07.790975225Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:42:22.542693436Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":340,"events_lines":2,"console_offset":166,"console_lines":1}
+{"time":"2026-07-08T05:42:22.820888218Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:42:37.54208972Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4224,"history_lines":1,"events_offset":342,"events_lines":2,"console_offset":167,"console_lines":1}
+{"time":"2026-07-08T05:42:37.871099747Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:42:52.54240899Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4225,"history_lines":1,"events_offset":344,"events_lines":2,"console_offset":168,"console_lines":1}
+{"time":"2026-07-08T05:42:52.895591482Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:43:07.542086605Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4226,"history_lines":2,"events_offset":346,"events_lines":2,"console_offset":169,"console_lines":2}
+{"time":"2026-07-08T05:43:07.893177882Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:43:22.542460509Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4228,"history_lines":1,"events_offset":348,"events_lines":2,"console_offset":171,"console_lines":1}
+{"time":"2026-07-08T05:43:22.802173228Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:43:37.542569448Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4229,"history_lines":1,"events_offset":350,"events_lines":2,"console_offset":172,"console_lines":1}
+{"time":"2026-07-08T05:43:37.919599519Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:43:52.542505729Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4230,"history_lines":1,"events_offset":352,"events_lines":2,"console_offset":173,"console_lines":1}
+{"time":"2026-07-08T05:43:52.835953984Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:44:07.541957967Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4231,"history_lines":1,"events_offset":354,"events_lines":2,"console_offset":174,"console_lines":1}
+{"time":"2026-07-08T05:44:07.881436771Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:44:22.542373124Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4232,"history_lines":1,"events_offset":356,"events_lines":2,"console_offset":175,"console_lines":1}
+{"time":"2026-07-08T05:44:22.83145214Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:44:37.54260008Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4233,"history_lines":1,"events_offset":358,"events_lines":2,"console_offset":176,"console_lines":1}
+{"time":"2026-07-08T05:44:38.064052639Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:44:52.541996545Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4234,"history_lines":1,"events_offset":360,"events_lines":2,"console_offset":177,"console_lines":1}
+{"time":"2026-07-08T05:44:52.865233396Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:45:07.541831645Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4235,"history_lines":1,"events_offset":362,"events_lines":2,"console_offset":178,"console_lines":1}
+{"time":"2026-07-08T05:45:07.943760595Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:45:22.541846314Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4236,"history_lines":1,"events_offset":364,"events_lines":2,"console_offset":179,"console_lines":1}
+{"time":"2026-07-08T05:45:22.878857309Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:45:37.542731883Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4237,"history_lines":1,"events_offset":366,"events_lines":2,"console_offset":180,"console_lines":1}
+{"time":"2026-07-08T05:45:37.933390821Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:45:52.542492773Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4238,"history_lines":1,"events_offset":368,"events_lines":2,"console_offset":181,"console_lines":1}
+{"time":"2026-07-08T05:45:52.88372777Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:46:07.542809272Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4239,"history_lines":1,"events_offset":370,"events_lines":2,"console_offset":182,"console_lines":1}
+{"time":"2026-07-08T05:46:07.87249475Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:46:22.542247027Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4240,"history_lines":2,"events_offset":372,"events_lines":2,"console_offset":183,"console_lines":2}
+{"time":"2026-07-08T05:46:22.859432124Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:46:37.542552931Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4242,"history_lines":1,"events_offset":374,"events_lines":2,"console_offset":185,"console_lines":1}
+{"time":"2026-07-08T05:46:37.900039715Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:46:52.542323234Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4243,"history_lines":1,"events_offset":376,"events_lines":2,"console_offset":186,"console_lines":1}
+{"time":"2026-07-08T05:46:52.890801126Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:47:07.542003993Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4244,"history_lines":1,"events_offset":378,"events_lines":2,"console_offset":187,"console_lines":1}
+{"time":"2026-07-08T05:47:07.886899932Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:47:22.542704828Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4245,"history_lines":1,"events_offset":380,"events_lines":2,"console_offset":188,"console_lines":1}
+{"time":"2026-07-08T05:47:22.855366645Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:47:37.542834318Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4246,"history_lines":1,"events_offset":382,"events_lines":2,"console_offset":189,"console_lines":1}
+{"time":"2026-07-08T05:47:37.880379051Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:47:52.542605803Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4247,"history_lines":1,"events_offset":384,"events_lines":2,"console_offset":190,"console_lines":1}
+{"time":"2026-07-08T05:47:53.059199089Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:48:07.542017891Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4248,"history_lines":1,"events_offset":386,"events_lines":2,"console_offset":191,"console_lines":1}
+{"time":"2026-07-08T05:48:08.048873826Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:48:22.54249489Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":388,"events_lines":2,"console_offset":192,"console_lines":1}
+{"time":"2026-07-08T05:48:22.780969701Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:48:37.542932541Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4249,"history_lines":1,"events_offset":390,"events_lines":2,"console_offset":193,"console_lines":1}
+{"time":"2026-07-08T05:48:38.051603498Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:48:52.542189637Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4250,"history_lines":1,"events_offset":392,"events_lines":2,"console_offset":194,"console_lines":1}
+{"time":"2026-07-08T05:48:53.114811604Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:49:07.542461921Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4251,"history_lines":1,"events_offset":394,"events_lines":2,"console_offset":195,"console_lines":1}
+{"time":"2026-07-08T05:49:07.831476399Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:49:22.542532183Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4252,"history_lines":1,"events_offset":396,"events_lines":2,"console_offset":196,"console_lines":1}
+{"time":"2026-07-08T05:49:22.84166323Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:49:37.542570336Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4253,"history_lines":1,"events_offset":398,"events_lines":2,"console_offset":197,"console_lines":1}
+{"time":"2026-07-08T05:49:37.809292731Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:49:52.542443924Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4254,"history_lines":1,"events_offset":400,"events_lines":2,"console_offset":198,"console_lines":1}
+{"time":"2026-07-08T05:49:52.928962745Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:50:07.542503009Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4255,"history_lines":1,"events_offset":402,"events_lines":2,"console_offset":199,"console_lines":1}
+{"time":"2026-07-08T05:50:07.949408997Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:50:22.541860085Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4256,"history_lines":1,"events_offset":404,"events_lines":2,"console_offset":200,"console_lines":1}
+{"time":"2026-07-08T05:50:22.868658668Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:50:37.542577964Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4257,"history_lines":1,"events_offset":406,"events_lines":2,"console_offset":201,"console_lines":1}
+{"time":"2026-07-08T05:50:37.90988189Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:50:52.542439313Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4258,"history_lines":1,"events_offset":408,"events_lines":2,"console_offset":202,"console_lines":1}
+{"time":"2026-07-08T05:50:52.855892463Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:51:07.542530807Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4259,"history_lines":2,"events_offset":410,"events_lines":2,"console_offset":203,"console_lines":2}
+{"time":"2026-07-08T05:51:07.869013869Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:51:22.542672275Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4261,"history_lines":1,"events_offset":412,"events_lines":2,"console_offset":205,"console_lines":1}
+{"time":"2026-07-08T05:51:22.869628795Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:51:37.542771139Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4262,"history_lines":1,"events_offset":414,"events_lines":2,"console_offset":206,"console_lines":1}
+{"time":"2026-07-08T05:51:37.849606269Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:51:52.542521212Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4263,"history_lines":1,"events_offset":416,"events_lines":2,"console_offset":207,"console_lines":1}
+{"time":"2026-07-08T05:51:52.89717287Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:52:07.54286357Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4264,"history_lines":1,"events_offset":418,"events_lines":2,"console_offset":208,"console_lines":1}
+{"time":"2026-07-08T05:52:07.851627563Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:52:22.5425919Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4265,"history_lines":1,"events_offset":420,"events_lines":2,"console_offset":209,"console_lines":1}
+{"time":"2026-07-08T05:52:22.995276791Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:52:37.542415022Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4266,"history_lines":1,"events_offset":422,"events_lines":2,"console_offset":210,"console_lines":1}
+{"time":"2026-07-08T05:52:37.847981807Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:52:52.542440858Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4267,"history_lines":1,"events_offset":424,"events_lines":2,"console_offset":211,"console_lines":1}
+{"time":"2026-07-08T05:52:53.062855295Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:53:07.542325762Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4268,"history_lines":1,"events_offset":426,"events_lines":2,"console_offset":212,"console_lines":1}
+{"time":"2026-07-08T05:53:07.957277083Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:53:22.542763603Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4269,"history_lines":1,"events_offset":428,"events_lines":2,"console_offset":213,"console_lines":1}
+{"time":"2026-07-08T05:53:22.869874474Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:53:37.542171144Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4270,"history_lines":1,"events_offset":430,"events_lines":2,"console_offset":214,"console_lines":1}
+{"time":"2026-07-08T05:53:37.846248493Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:53:52.544835112Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4271,"history_lines":1,"events_offset":432,"events_lines":2,"console_offset":215,"console_lines":1}
+{"time":"2026-07-08T05:53:52.88847943Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:54:07.542436334Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4272,"history_lines":1,"events_offset":434,"events_lines":2,"console_offset":216,"console_lines":1}
+{"time":"2026-07-08T05:54:07.770153919Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:54:22.542093107Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4273,"history_lines":1,"events_offset":436,"events_lines":2,"console_offset":217,"console_lines":1}
+{"time":"2026-07-08T05:54:22.824515745Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:54:37.541939885Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":438,"events_lines":2,"console_offset":218,"console_lines":1}
+{"time":"2026-07-08T05:54:37.781186089Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:54:52.542680719Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4274,"history_lines":1,"events_offset":440,"events_lines":2,"console_offset":219,"console_lines":1}
+{"time":"2026-07-08T05:54:52.879867196Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:55:07.541760691Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4275,"history_lines":1,"events_offset":442,"events_lines":2,"console_offset":220,"console_lines":1}
+{"time":"2026-07-08T05:55:07.83521246Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:55:22.54190339Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4276,"history_lines":1,"events_offset":444,"events_lines":2,"console_offset":221,"console_lines":1}
+{"time":"2026-07-08T05:55:23.131566574Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:55:37.542318848Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4277,"history_lines":1,"events_offset":446,"events_lines":2,"console_offset":222,"console_lines":1}
+{"time":"2026-07-08T05:55:37.859192239Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:55:52.542034439Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4278,"history_lines":1,"events_offset":448,"events_lines":2,"console_offset":223,"console_lines":1}
+{"time":"2026-07-08T05:55:53.118235825Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:56:07.542145221Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4279,"history_lines":1,"events_offset":450,"events_lines":2,"console_offset":224,"console_lines":1}
+{"time":"2026-07-08T05:56:07.894920856Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:56:22.542086971Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4280,"history_lines":2,"events_offset":452,"events_lines":2,"console_offset":225,"console_lines":2}
+{"time":"2026-07-08T05:56:23.044631953Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:56:37.542177983Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4282,"history_lines":1,"events_offset":454,"events_lines":2,"console_offset":227,"console_lines":1}
+{"time":"2026-07-08T05:56:37.949087016Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:56:52.542663055Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4283,"history_lines":1,"events_offset":456,"events_lines":2,"console_offset":228,"console_lines":1}
+{"time":"2026-07-08T05:56:52.955769551Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:57:07.542264131Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4284,"history_lines":1,"events_offset":458,"events_lines":2,"console_offset":229,"console_lines":1}
+{"time":"2026-07-08T05:57:08.049030572Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:57:22.542792012Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4285,"history_lines":1,"events_offset":460,"events_lines":2,"console_offset":230,"console_lines":1}
+{"time":"2026-07-08T05:57:22.812125073Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:57:37.542207922Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4286,"history_lines":1,"events_offset":462,"events_lines":2,"console_offset":231,"console_lines":1}
+{"time":"2026-07-08T05:57:37.877971867Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:57:52.542624787Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4287,"history_lines":1,"events_offset":464,"events_lines":2,"console_offset":232,"console_lines":1}
+{"time":"2026-07-08T05:57:53.200296851Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:58:07.54245823Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4288,"history_lines":1,"events_offset":466,"events_lines":2,"console_offset":233,"console_lines":1}
+{"time":"2026-07-08T05:58:07.925113379Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}

epfl_final/SparseAtt/wandb/run-20260708_045920-qlqhc3gy/logs/debug.log

@@ -0,0 +1,20 @@
+2026-07-08 04:59:20,853 INFO    MainThread:575 [wandb_setup.py:_flush():81] Current SDK version is 0.26.1
+2026-07-08 04:59:20,854 INFO    MainThread:575 [wandb_setup.py:_flush():81] Configure stats pid to 575
+2026-07-08 04:59:20,854 INFO    MainThread:575 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2026-07-08 04:59:20,854 INFO    MainThread:575 [wandb_init.py:setup_run_log_directory():723] Logging user logs to /mloscratch/homes/navasard/sparse/SparseAtt/wandb/run-20260708_045920-qlqhc3gy/logs/debug.log
+2026-07-08 04:59:20,854 INFO    MainThread:575 [wandb_init.py:setup_run_log_directory():724] Logging internal logs to /mloscratch/homes/navasard/sparse/SparseAtt/wandb/run-20260708_045920-qlqhc3gy/logs/debug-internal.log
+2026-07-08 04:59:20,854 INFO    MainThread:575 [wandb_init.py:init():850] calling init triggers
+2026-07-08 04:59:20,854 INFO    MainThread:575 [wandb_init.py:init():855] wandb.init called with sweep_config: {}
+config: {'model_path': '/mloscratch/homes/navasard/qwen3_06b_base', 'model_dtype': 'bfloat16', 'init_from_scratch': True, 'train_path': '/mloscratch/homes/navasard/data/fineweb_large/tokenized', 'val_path': None, 'max_seq_length': 8192, 'num_workers': 8, 'output_dir': '/mloscratch/homes/navasard/sparse/checkpoints/l8k365m_dense/run_20260707_034553', 'resume_from_checkpoint': '/mloscratch/homes/navasard/sparse/checkpoints/l8k365m_dense/run_20260707_034553/checkpoint-41000', 'force_new_run': False, 'num_epochs': 1, 'max_steps': 44000, 'per_device_batch_size': 2, 'gradient_accumulation_steps': 10, 'learning_rate': 0.0005, 'weight_decay': 0.01, 'warmup_steps': 750, 'lr_scheduler': 'cosine', 'max_grad_norm': 1.0, 'save_steps': 1000, 'eval_steps': 250, 'eval_batches': 50, 'logging_steps': 10, 'seed': 42, 'verbose': True, 'checkpoint_activations': False, 'strategy': 'ddp', 'mixed_precision': 'bf16', 'wandb_enabled': True, 'wandb_project': 'stochastic_sparsity', 'wandb_name': '365m dense 8k', 'wandb_tags': ['qwen3', '365m', '8k', 'scaleladder', 'dense'], 'hidden_size': 768, 'intermediate_size': 3072, 'num_attention_heads': 6, 'num_key_value_heads': 2, 'num_hidden_layers': 28, 'head_dim': None, 'attention_type': 'sparse_stochastic', 'sparse_block_size': 128, 'sparse_num_random_blocks': 31, 'sparse_mask_pool_size': 1, 'sparse_val_dense_compare': False, 'sparse_layer_indices': None, 'sparse_mask_cache_dir': None, 'sparse_mask_seed': 0, 'dense_mode': True, 'dense_step_intervals': None, '_wandb': {}}
+2026-07-08 04:59:20,854 INFO    MainThread:575 [wandb_init.py:init():898] starting backend
+2026-07-08 04:59:21,416 INFO    MainThread:575 [wandb_init.py:init():913] sending inform_init request
+2026-07-08 04:59:21,860 INFO    MainThread:575 [wandb_init.py:init():918] backend started and connected
+2026-07-08 04:59:21,866 INFO    MainThread:575 [wandb_init.py:init():988] updated telemetry
+2026-07-08 04:59:21,897 INFO    MainThread:575 [wandb_init.py:init():1011] communicating run to backend with 90.0 second timeout
+2026-07-08 04:59:22,537 INFO    MainThread:575 [wandb_init.py:init():1051] run resumed
+2026-07-08 04:59:22,540 INFO    MainThread:575 [wandb_init.py:init():1056] starting run threads in backend
+2026-07-08 04:59:22,810 INFO    MainThread:575 [wandb_run.py:_console_start():2554] atexit reg
+2026-07-08 04:59:22,810 INFO    MainThread:575 [wandb_run.py:_redirect():2403] redirect: wrap_raw
+2026-07-08 04:59:22,810 INFO    MainThread:575 [wandb_run.py:_redirect():2472] Wrapping output streams.
+2026-07-08 04:59:22,811 INFO    MainThread:575 [wandb_run.py:_redirect():2495] Redirects installed.
+2026-07-08 04:59:22,820 INFO    MainThread:575 [wandb_init.py:init():1094] run started, returning control to user process

epfl_final/SparseAtt/wandb/run-20260708_045920-qlqhc3gy/run-qlqhc3gy.wandb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:065f6a60dfb0fbd1655d44e349ff3cbccd48ed300e9f109400152844fe9e49ad
+size 622592

epfl_final/SparseAtt/wandb/run-20260708_054405-43sa4ouk/files/output.log

@@ -0,0 +1,37 @@
+Starting training...
+Switching attention mode for step 1: dense (dense step interval)
+Step 10/37465 | Loss: 12.2534 | LR: 6.67e-06 | 16167.7ms/step | 1.2 samples/s
+Step 20/37465 | Loss: 12.1201 | LR: 1.33e-05 | 6304.5ms/step | 3.2 samples/s
+Step 30/37465 | Loss: 11.6858 | LR: 2.00e-05 | 6309.1ms/step | 3.2 samples/s
+Step 40/37465 | Loss: 10.6684 | LR: 2.67e-05 | 6298.4ms/step | 3.2 samples/s
+Step 50/37465 | Loss: 10.0551 | LR: 3.33e-05 | 6286.7ms/step | 3.2 samples/s
+Step 60/37465 | Loss: 9.5722 | LR: 4.00e-05 | 6268.7ms/step | 3.2 samples/s
+Step 70/37465 | Loss: 9.1242 | LR: 4.67e-05 | 6271.1ms/step | 3.2 samples/s
+Step 80/37465 | Loss: 8.6456 | LR: 5.33e-05 | 6278.9ms/step | 3.2 samples/s
+Step 90/37465 | Loss: 8.0658 | LR: 6.00e-05 | 6281.1ms/step | 3.2 samples/s
+Step 100/37465 | Loss: 7.7196 | LR: 6.67e-05 | 6269.8ms/step | 3.2 samples/s
+Step 110/37465 | Loss: 7.3699 | LR: 7.33e-05 | 6276.4ms/step | 3.2 samples/s
+Step 120/37465 | Loss: 7.1081 | LR: 8.00e-05 | 6290.5ms/step | 3.2 samples/s
+Step 130/37465 | Loss: 6.9196 | LR: 8.67e-05 | 6302.0ms/step | 3.2 samples/s
+Step 140/37465 | Loss: 6.8029 | LR: 9.33e-05 | 6285.6ms/step | 3.2 samples/s
+Step 150/37465 | Loss: 6.8077 | LR: 1.00e-04 | 6286.5ms/step | 3.2 samples/s
+Step 160/37465 | Loss: 6.6545 | LR: 1.07e-04 | 6303.8ms/step | 3.2 samples/s
+Step 170/37465 | Loss: 6.3287 | LR: 1.13e-04 | 6301.6ms/step | 3.2 samples/s
+Step 180/37465 | Loss: 6.1895 | LR: 1.20e-04 | 6295.4ms/step | 3.2 samples/s
+Step 190/37465 | Loss: 6.2694 | LR: 1.27e-04 | 6300.7ms/step | 3.2 samples/s
+Step 200/37465 | Loss: 6.0386 | LR: 1.33e-04 | 6304.3ms/step | 3.2 samples/s
+Step 210/37465 | Loss: 6.0358 | LR: 1.40e-04 | 6303.8ms/step | 3.2 samples/s
+Step 220/37465 | Loss: 5.9696 | LR: 1.47e-04 | 6309.1ms/step | 3.2 samples/s
+Step 230/37465 | Loss: 5.9117 | LR: 1.53e-04 | 6327.0ms/step | 3.2 samples/s
+Step 240/37465 | Loss: 5.8905 | LR: 1.60e-04 | 6389.6ms/step | 3.1 samples/s
+Step 250/37465 | Loss: 5.8312 | LR: 1.67e-04 | 6367.4ms/step | 3.1 samples/s
+Step 260/37465 | Loss: 5.6800 | LR: 1.73e-04 | 6363.2ms/step | 3.1 samples/s
+Step 270/37465 | Loss: 5.7001 | LR: 1.80e-04 | 6299.8ms/step | 3.2 samples/s
+Step 280/37465 | Loss: 5.7136 | LR: 1.87e-04 | 6306.0ms/step | 3.2 samples/s
+Step 290/37465 | Loss: 5.4759 | LR: 1.93e-04 | 6300.9ms/step | 3.2 samples/s
+Step 300/37465 | Loss: 5.3876 | LR: 2.00e-04 | 6309.8ms/step | 3.2 samples/s
+Step 310/37465 | Loss: 5.2540 | LR: 2.07e-04 | 6324.4ms/step | 3.2 samples/s
+Step 320/37465 | Loss: 5.3460 | LR: 2.13e-04 | 6306.3ms/step | 3.2 samples/s
+Step 330/37465 | Loss: 5.5442 | LR: 2.20e-04 | 6306.8ms/step | 3.2 samples/s
+Step 340/37465 | Loss: 5.3776 | LR: 2.27e-04 | 6312.8ms/step | 3.2 samples/s
+Step 350/37465 | Loss: 5.2115 | LR: 2.33e-04 | 6318.3ms/step | 3.2 samples/s

epfl_final/SparseAtt/wandb/run-20260708_054405-43sa4ouk/files/requirements.txt

@@ -0,0 +1,141 @@
+pip==26.0.1
+nvidia-cusparselt-cu13==0.8.0
+mpmath==1.3.0
+cuda-toolkit==13.0.2
+zstandard==0.25.0
+xxhash==3.7.0
+urllib3==2.6.3
+typing_extensions==4.15.0
+triton==3.6.0
+tqdm==4.67.3
+torchinfo==1.8.0
+sympy==1.14.0
+smmap==5.0.3
+six==1.17.0
+shellingham==1.5.4
+setuptools==81.0.0
+safetensors==0.7.0
+regex==2026.4.4
+PyYAML==6.0.3
+Pygments==2.20.0
+pyarrow==24.0.0
+psutil==7.2.2
+protobuf==7.34.1
+propcache==0.4.1
+platformdirs==4.9.6
+packaging==26.2
+nvidia-nvtx==13.0.85
+nvidia-nvshmem-cu13==3.4.5
+nvidia-nvjitlink==13.0.88
+nvidia-nccl-cu13==2.28.9
+nvidia-curand==10.4.0.35
+nvidia-cufile==1.15.1.6
+nvidia-cuda-runtime==13.0.96
+nvidia-cuda-nvrtc==13.0.88
+nvidia-cuda-cupti==13.0.85
+nvidia-cublas==13.1.0.3
+numpy==2.4.4
+networkx==3.6.1
+multidict==6.7.1
+mdurl==0.1.2
+MarkupSafe==3.0.3
+idna==3.13
+hf-xet==1.4.3
+h11==0.16.0
+fsspec==2026.2.0
+frozenlist==1.8.0
+filelock==3.29.0
+dill==0.4.1
+cuda-pathfinder==1.5.4
+click==8.3.3
+charset-normalizer==3.4.7
+certifi==2026.4.22
+attrs==26.1.0
+annotated-types==0.7.0
+annotated-doc==0.0.4
+aiohappyeyeballs==2.6.1
+yarl==1.23.0
+typing-inspection==0.4.2
+sentry-sdk==2.58.0
+requests==2.33.1
+python-dateutil==2.9.0.post0
+pydantic_core==2.46.3
+nvidia-cusparse==12.6.3.3
+nvidia-cufft==12.0.0.61
+nvidia-cudnn-cu13==9.19.0.56
+multiprocess==0.70.19
+markdown-it-py==4.0.0
+Jinja2==3.1.6
+httpcore==1.0.9
+gitdb==4.0.12
+cuda-bindings==13.2.0
+anyio==4.13.0
+aiosignal==1.4.0
+rich==15.0.0
+pydantic==2.13.3
+pandas==3.0.2
+nvidia-cusolver==12.0.4.66
+httpx==0.28.1
+GitPython==3.1.49
+aiohttp==3.13.5
+wandb==0.26.1
+typer==0.25.1
+torch==2.11.0
+huggingface_hub==1.13.0
+tokenizers==0.22.2
+datasets==4.8.5
+accelerate==1.13.0
+transformers==5.7.0
+word2number==1.1
+sqlitedict==2.1.0
+pytz==2026.2
+threadpoolctl==3.6.0
+tcolorpy==0.1.7
+tabulate==0.10.0
+scipy==1.17.1
+portalocker==3.2.0
+pathvalidate==3.3.1
+narwhals==2.22.1
+more-itertools==11.1.0
+lxml==6.1.1
+joblib==1.5.3
+colorama==0.4.6
+chardet==6.0.0.post1
+absl-py==2.4.0
+scikit-learn==1.9.0
+sacrebleu==2.6.0
+nltk==3.9.4
+mbstrdecoder==1.1.5
+typepy==1.3.5
+rouge_score==0.1.2
+DataProperty==1.1.1
+tabledata==1.3.5
+evaluate==0.4.6
+pytablewriter==1.2.1
+lm_eval==0.4.12
+pluggy==1.6.0
+iniconfig==2.3.0
+pytest==9.1.0
+py-spy==0.4.2
+pillow==12.2.0
+tifffile==2026.6.1
+imagecodecs==2026.6.26
+zipp==4.1.0
+importlib_metadata==9.0.0
+derivative==0.6.3
+pysindy==2.1.0
+pyparsing==3.3.2
+kiwisolver==1.5.0
+fonttools==4.63.0
+cycler==0.12.1
+contourpy==1.3.3
+matplotlib==3.11.0
+sortedcontainers==2.4.0
+wcwidth==0.8.2
+sigfig==1.3.19
+importlib_resources==7.1.0
+enum-tools==0.13.0
+cramjam==2.11.0
+python-snappy==0.7.3
+compact-json==1.8.2
+numbers-parser==4.18.5

epfl_final/SparseAtt/wandb/run-20260708_054405-43sa4ouk/files/wandb-metadata.json

@@ -0,0 +1,65 @@
+{
+  "os":  "Linux-6.8.0-124-generic-x86_64-with-glibc2.35",
+  "python":  "CPython 3.13.13",
+  "startedAt":  "2026-07-08T05:44:05.569383Z",
+  "args":  [
+    "--config",
+    "configs/train_config_30k_dense_b128.json"
+  ],
+  "program":  "/mloscratch/homes/navasard/sparse/SparseAtt/train.py",
+  "codePath":  "train.py",
+  "codePathLocal":  "train.py",
+  "git":  {
+    "remote":  "https://github.com/rubentium/TokenPrune.git",
+    "commit":  "88298c7a418622530ca5426b22b1e5ebd10ca0cb"
+  },
+  "email":  "ruben.navasardyan@mail.utoronto.ca",
+  "root":  "/mloscratch/homes/navasard/sparse/SparseAtt",
+  "host":  "h100-train-4-gpu-2-0-1",
+  "executable":  "/mloscratch/homes/navasard/conda/bin/python3.13",
+  "cpu_count":  96,
+  "cpu_count_logical":  192,
+  "gpu":  "NVIDIA H100 80GB HBM3",
+  "gpu_count":  4,
+  "disk":  {
+    "/":  {
+      "total":  "9371709865984",
+      "used":  "5833919746048"
+    }
+  },
+  "memory":  {
+    "total":  "1623005732864"
+  },
+  "gpu_nvidia":  [
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-98029aa4-c589-b65b-3ab7-6495256f31ed"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-37a1c6c2-d3df-3064-44df-33fa82304bd0"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-b87e8111-63b6-b364-28b9-c4ce747aec08"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-bbbaaf81-9234-a1c6-0093-6d87b4e6af4c"
+    }
+  ],
+  "cudaVersion":  "13.0",
+  "writerId":  "y70t0m028ds3frs3dcd4sibkpw6ly3z9"
+}

epfl_final/SparseAtt/wandb/run-20260708_054405-43sa4ouk/logs/debug-internal.log

@@ -0,0 +1,321 @@
+{"time":"2026-07-08T05:44:06.142860826Z","level":"INFO","msg":"wandb-core"}
+{"time":"2026-07-08T05:44:06.147642849Z","level":"INFO","msg":"stream: starting","core version":"0.26.1"}
+{"time":"2026-07-08T05:44:06.647939183Z","level":"INFO","msg":"stream: created new stream","id":"43sa4ouk"}
+{"time":"2026-07-08T05:44:06.648066854Z","level":"INFO","msg":"handler: started"}
+{"time":"2026-07-08T05:44:06.64988514Z","level":"INFO","msg":"stream: started"}
+{"time":"2026-07-08T05:44:06.650827047Z","level":"INFO","msg":"sender: started"}
+{"time":"2026-07-08T05:44:06.650831223Z","level":"INFO","msg":"writer: started","stream_id":"43sa4ouk"}
+{"time":"2026-07-08T05:44:07.323393762Z","level":"INFO","msg":"filestream: sending request","total_files":1,"console_offset":0,"console_lines":1}
+{"time":"2026-07-08T05:44:07.672285745Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:44:22.323509618Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":0,"events_lines":2,"console_offset":1,"console_lines":1,"uploaded_len":2}
+{"time":"2026-07-08T05:44:22.658465954Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:44:37.323677882Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":2,"events_lines":2}
+{"time":"2026-07-08T05:44:37.650998942Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:44:52.324286053Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":4,"events_lines":2}
+{"time":"2026-07-08T05:44:52.655533698Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:45:07.324226236Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":6,"events_lines":2}
+{"time":"2026-07-08T05:45:07.614784939Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:45:22.32339714Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":8,"events_lines":2}
+{"time":"2026-07-08T05:45:22.652044471Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:45:37.324121675Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":10,"events_lines":2}
+{"time":"2026-07-08T05:45:37.627599197Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:45:52.323599768Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":12,"events_lines":2}
+{"time":"2026-07-08T05:45:52.747788761Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:46:07.324427116Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":14,"events_lines":2}
+{"time":"2026-07-08T05:46:07.64642928Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:46:22.323923035Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":16,"events_lines":2}
+{"time":"2026-07-08T05:46:22.757604258Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:46:37.323739784Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":18,"events_lines":2}
+{"time":"2026-07-08T05:46:37.75163258Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:46:52.323777603Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":20,"events_lines":2,"console_offset":2,"console_lines":1}
+{"time":"2026-07-08T05:46:52.67537664Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:47:07.323846468Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":22,"events_lines":2}
+{"time":"2026-07-08T05:47:07.65977091Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:47:22.324158026Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":24,"events_lines":2}
+{"time":"2026-07-08T05:47:22.692477861Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:47:37.323418664Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":26,"events_lines":2}
+{"time":"2026-07-08T05:47:37.633785287Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:47:52.323914127Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":0,"history_lines":1,"events_offset":28,"events_lines":2,"console_offset":3,"console_lines":1}
+{"time":"2026-07-08T05:47:52.648956613Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:48:07.324379194Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":30,"events_lines":2}
+{"time":"2026-07-08T05:48:07.601279164Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:48:22.32422097Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":32,"events_lines":2}
+{"time":"2026-07-08T05:48:22.653600835Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:48:37.323906152Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":34,"events_lines":2}
+{"time":"2026-07-08T05:48:37.700233248Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:48:52.324396818Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":36,"events_lines":2}
+{"time":"2026-07-08T05:48:52.629029755Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:49:07.323863253Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":1,"history_lines":1,"events_offset":38,"events_lines":2,"console_offset":4,"console_lines":1}
+{"time":"2026-07-08T05:49:07.626331097Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:49:22.32421417Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":40,"events_lines":2}
+{"time":"2026-07-08T05:49:22.563845697Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:49:37.323625914Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":42,"events_lines":2}
+{"time":"2026-07-08T05:49:37.759402557Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:49:52.323676592Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":44,"events_lines":2}
+{"time":"2026-07-08T05:49:52.814324093Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:50:07.323625549Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":2,"history_lines":1,"events_offset":46,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-07-08T05:50:07.972535866Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:50:22.324278126Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":48,"events_lines":2}
+{"time":"2026-07-08T05:50:22.672895549Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:50:37.323632153Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":50,"events_lines":2}
+{"time":"2026-07-08T05:50:37.690163456Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:50:52.324261376Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":52,"events_lines":2}
+{"time":"2026-07-08T05:50:52.68128777Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:51:07.324466454Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":3,"history_lines":1,"events_offset":54,"events_lines":2,"console_offset":6,"console_lines":1}
+{"time":"2026-07-08T05:51:07.856323084Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:51:22.323767723Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":56,"events_lines":2}
+{"time":"2026-07-08T05:51:22.638048063Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:51:37.323829017Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":58,"events_lines":2}
+{"time":"2026-07-08T05:51:37.67860496Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:51:52.323993085Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":60,"events_lines":2}
+{"time":"2026-07-08T05:51:52.663752085Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:52:07.324413696Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4,"history_lines":1,"events_offset":62,"events_lines":2,"console_offset":7,"console_lines":1}
+{"time":"2026-07-08T05:52:07.615545475Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:52:22.323977456Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":64,"events_lines":2}
+{"time":"2026-07-08T05:52:22.745791622Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:52:37.323772232Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":66,"events_lines":2}
+{"time":"2026-07-08T05:52:37.747250534Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:52:52.323672606Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":68,"events_lines":2}
+{"time":"2026-07-08T05:52:52.768262234Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:53:07.323915832Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":5,"history_lines":1,"events_offset":70,"events_lines":2,"console_offset":8,"console_lines":1}
+{"time":"2026-07-08T05:53:07.552316832Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:53:22.324157295Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":72,"events_lines":2}
+{"time":"2026-07-08T05:53:22.737320427Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:53:37.324388693Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":74,"events_lines":2}
+{"time":"2026-07-08T05:53:37.737974216Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:53:52.323752134Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":76,"events_lines":2}
+{"time":"2026-07-08T05:53:52.689070175Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:54:07.323682904Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":78,"events_lines":2}
+{"time":"2026-07-08T05:54:07.625435639Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:54:22.323840051Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":6,"history_lines":1,"events_offset":80,"events_lines":2,"console_offset":9,"console_lines":1}
+{"time":"2026-07-08T05:54:22.680965042Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:54:37.324352409Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":82,"events_lines":2}
+{"time":"2026-07-08T05:54:37.817242499Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:54:52.323805304Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":84,"events_lines":2}
+{"time":"2026-07-08T05:54:52.673049863Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:55:07.324074698Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":86,"events_lines":2}
+{"time":"2026-07-08T05:55:07.612797929Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:55:22.32389345Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":7,"history_lines":1,"events_offset":88,"events_lines":2,"console_offset":10,"console_lines":1}
+{"time":"2026-07-08T05:55:22.697795353Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:55:37.324524996Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":90,"events_lines":2}
+{"time":"2026-07-08T05:55:37.709095366Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:55:52.324047865Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":92,"events_lines":2}
+{"time":"2026-07-08T05:55:52.886482423Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:56:07.323975522Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":94,"events_lines":2}
+{"time":"2026-07-08T05:56:07.690773667Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:56:22.324410699Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":8,"history_lines":1,"events_offset":96,"events_lines":2,"console_offset":11,"console_lines":1}
+{"time":"2026-07-08T05:56:22.676196593Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:56:37.324011991Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":98,"events_lines":2}
+{"time":"2026-07-08T05:56:37.721888652Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:56:52.324131457Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":100,"events_lines":2}
+{"time":"2026-07-08T05:56:52.585898207Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:57:07.32383352Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":102,"events_lines":2}
+{"time":"2026-07-08T05:57:07.743430751Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:57:22.324012234Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":9,"history_lines":1,"events_offset":104,"events_lines":2,"console_offset":12,"console_lines":1}
+{"time":"2026-07-08T05:57:22.690110835Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:57:37.324204458Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":106,"events_lines":2}
+{"time":"2026-07-08T05:57:37.628689178Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:57:52.324289803Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":108,"events_lines":2}
+{"time":"2026-07-08T05:57:53.166363756Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:58:07.323830385Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":110,"events_lines":2}
+{"time":"2026-07-08T05:58:07.976337578Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:58:22.324548274Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":10,"history_lines":1,"events_offset":112,"events_lines":2,"console_offset":13,"console_lines":1}
+{"time":"2026-07-08T05:58:22.803986498Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:58:37.324135065Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":114,"events_lines":2}
+{"time":"2026-07-08T05:58:37.759460536Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:58:52.324066259Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":116,"events_lines":2}
+{"time":"2026-07-08T05:58:52.661408632Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:59:07.324178955Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":118,"events_lines":2}
+{"time":"2026-07-08T05:59:07.778872553Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:59:22.324383448Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":120,"events_lines":2}
+{"time":"2026-07-08T05:59:22.711606218Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:59:37.323834626Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":11,"history_lines":1,"events_offset":122,"events_lines":2,"console_offset":14,"console_lines":1}
+{"time":"2026-07-08T05:59:37.72361455Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T05:59:52.323804849Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":124,"events_lines":2}
+{"time":"2026-07-08T05:59:52.804836974Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:00:07.323801511Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":126,"events_lines":2}
+{"time":"2026-07-08T06:00:07.646743437Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:00:22.323733668Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":128,"events_lines":2}
+{"time":"2026-07-08T06:00:22.754296046Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:00:37.324213402Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":12,"history_lines":1,"events_offset":130,"events_lines":2,"console_offset":15,"console_lines":1}
+{"time":"2026-07-08T06:00:37.675364919Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:00:52.324278777Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":132,"events_lines":2}
+{"time":"2026-07-08T06:00:52.656091031Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:01:07.323774461Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":134,"events_lines":2}
+{"time":"2026-07-08T06:01:07.660949925Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:01:22.324165062Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":136,"events_lines":2}
+{"time":"2026-07-08T06:01:22.667575177Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:01:37.324268594Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":13,"history_lines":1,"events_offset":138,"events_lines":2,"console_offset":16,"console_lines":1}
+{"time":"2026-07-08T06:01:37.622103893Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:01:52.32394613Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":140,"events_lines":2}
+{"time":"2026-07-08T06:01:52.644660349Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:02:07.323647151Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":142,"events_lines":2}
+{"time":"2026-07-08T06:02:07.719777735Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:02:22.324061347Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":144,"events_lines":2}
+{"time":"2026-07-08T06:02:22.653269181Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:02:37.324159881Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":14,"history_lines":1,"events_offset":146,"events_lines":2,"console_offset":17,"console_lines":1}
+{"time":"2026-07-08T06:02:37.56830273Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:02:52.323640764Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":148,"events_lines":2}
+{"time":"2026-07-08T06:02:52.595924802Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:03:07.324132956Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":150,"events_lines":2}
+{"time":"2026-07-08T06:03:07.792606525Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:03:22.324162007Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":152,"events_lines":2}
+{"time":"2026-07-08T06:03:22.768344587Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:03:37.324457015Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":15,"history_lines":1,"events_offset":154,"events_lines":2,"console_offset":18,"console_lines":1}
+{"time":"2026-07-08T06:03:37.784784547Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:03:52.323428026Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":156,"events_lines":2}
+{"time":"2026-07-08T06:03:52.764288794Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:04:07.323754411Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":158,"events_lines":2}
+{"time":"2026-07-08T06:04:07.804197255Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:04:22.323465237Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":160,"events_lines":2}
+{"time":"2026-07-08T06:04:22.723877355Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:04:37.324251268Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":162,"events_lines":2}
+{"time":"2026-07-08T06:04:37.692784595Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:04:52.323783406Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":16,"history_lines":1,"events_offset":164,"events_lines":2,"console_offset":19,"console_lines":1}
+{"time":"2026-07-08T06:04:52.766742689Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:05:07.324152996Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":166,"events_lines":2}
+{"time":"2026-07-08T06:05:07.574081956Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:05:22.32438534Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":168,"events_lines":2}
+{"time":"2026-07-08T06:05:22.826078242Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:05:37.324351978Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":170,"events_lines":2}
+{"time":"2026-07-08T06:05:37.628634726Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:05:52.32353106Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":17,"history_lines":1,"events_offset":172,"events_lines":2,"console_offset":20,"console_lines":1}
+{"time":"2026-07-08T06:05:52.684943822Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:06:07.323547522Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":174,"events_lines":2}
+{"time":"2026-07-08T06:06:07.678873423Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:06:22.324025222Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":176,"events_lines":2}
+{"time":"2026-07-08T06:06:22.646778416Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:06:37.323612374Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":178,"events_lines":2}
+{"time":"2026-07-08T06:06:37.595877523Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:06:52.324397313Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":18,"history_lines":1,"events_offset":180,"events_lines":2,"console_offset":21,"console_lines":1}
+{"time":"2026-07-08T06:06:52.688593362Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:07:07.323807289Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":182,"events_lines":2}
+{"time":"2026-07-08T06:07:07.660275207Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:07:22.324053034Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":184,"events_lines":2}
+{"time":"2026-07-08T06:07:22.632921071Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:07:37.324298629Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":186,"events_lines":2}
+{"time":"2026-07-08T06:07:37.643868138Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:07:52.324181511Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":19,"history_lines":1,"events_offset":188,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-07-08T06:07:52.727142646Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:08:07.323643485Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":190,"events_lines":2}
+{"time":"2026-07-08T06:08:07.839381483Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:08:22.32358105Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":192,"events_lines":2}
+{"time":"2026-07-08T06:08:22.580642012Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:08:37.323674867Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":194,"events_lines":2}
+{"time":"2026-07-08T06:08:37.594607677Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:08:52.324094171Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":20,"history_lines":1,"events_offset":196,"events_lines":2,"console_offset":23,"console_lines":1}
+{"time":"2026-07-08T06:08:52.710778657Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:09:07.323930894Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":198,"events_lines":2}
+{"time":"2026-07-08T06:09:07.690583643Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:09:22.323669221Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":200,"events_lines":2}
+{"time":"2026-07-08T06:09:22.747027029Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:09:37.323466126Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":202,"events_lines":2}
+{"time":"2026-07-08T06:09:37.722177073Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:09:52.32434761Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":204,"events_lines":2}
+{"time":"2026-07-08T06:09:52.756237901Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:10:07.3244834Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":21,"history_lines":1,"events_offset":206,"events_lines":2,"console_offset":24,"console_lines":1}
+{"time":"2026-07-08T06:10:07.770389013Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:10:22.323664495Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":208,"events_lines":2}
+{"time":"2026-07-08T06:10:22.752550559Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:10:37.32405792Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":210,"events_lines":2}
+{"time":"2026-07-08T06:10:37.825254049Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:10:52.3235242Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":212,"events_lines":2}
+{"time":"2026-07-08T06:10:52.573295014Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:11:07.323660741Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":22,"history_lines":1,"events_offset":214,"events_lines":2,"console_offset":25,"console_lines":1}
+{"time":"2026-07-08T06:11:07.854013787Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:11:22.323906596Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":216,"events_lines":2}
+{"time":"2026-07-08T06:11:22.676205827Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:11:37.32420525Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":218,"events_lines":2}
+{"time":"2026-07-08T06:11:37.608066919Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:11:52.323815575Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":220,"events_lines":2}
+{"time":"2026-07-08T06:11:52.730978235Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:12:07.324059737Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":23,"history_lines":1,"events_offset":222,"events_lines":2,"console_offset":26,"console_lines":1}
+{"time":"2026-07-08T06:12:07.792059408Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:12:22.32416346Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":224,"events_lines":2}
+{"time":"2026-07-08T06:12:22.692041408Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:12:37.324443165Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":226,"events_lines":2}
+{"time":"2026-07-08T06:12:37.553726551Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:12:52.323372323Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":228,"events_lines":2}
+{"time":"2026-07-08T06:12:52.661757223Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:13:07.323755Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":24,"history_lines":1,"events_offset":230,"events_lines":2,"console_offset":27,"console_lines":1}
+{"time":"2026-07-08T06:13:07.632399439Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:13:22.32339689Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":232,"events_lines":2}
+{"time":"2026-07-08T06:13:22.609340616Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:13:37.323381802Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":234,"events_lines":2}
+{"time":"2026-07-08T06:13:37.69045069Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:13:52.323532261Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":236,"events_lines":2}
+{"time":"2026-07-08T06:13:52.623180811Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:14:07.323425517Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":238,"events_lines":2}
+{"time":"2026-07-08T06:14:07.81332507Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:14:22.324472806Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":25,"history_lines":1,"events_offset":240,"events_lines":2,"console_offset":28,"console_lines":1}
+{"time":"2026-07-08T06:14:22.742096717Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:14:37.323938613Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":242,"events_lines":2}
+{"time":"2026-07-08T06:14:37.673403262Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:14:52.323396268Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":244,"events_lines":2}
+{"time":"2026-07-08T06:14:52.74117498Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:15:07.323589751Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":246,"events_lines":2}
+{"time":"2026-07-08T06:15:07.787323115Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:15:22.323771658Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":26,"history_lines":1,"events_offset":248,"events_lines":2,"console_offset":29,"console_lines":1}
+{"time":"2026-07-08T06:15:22.705786233Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:15:37.323822038Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":250,"events_lines":2}
+{"time":"2026-07-08T06:15:37.613834342Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:15:52.323783946Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":252,"events_lines":2}
+{"time":"2026-07-08T06:15:52.810934131Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:16:07.323577893Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":254,"events_lines":2}
+{"time":"2026-07-08T06:16:07.867759176Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:16:22.323725918Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":27,"history_lines":1,"events_offset":256,"events_lines":2,"console_offset":30,"console_lines":1}
+{"time":"2026-07-08T06:16:22.730368966Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:16:37.323856999Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":258,"events_lines":2}
+{"time":"2026-07-08T06:16:37.782084968Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:16:52.323398549Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":260,"events_lines":2}
+{"time":"2026-07-08T06:16:52.635901051Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:17:07.324180364Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":262,"events_lines":2}
+{"time":"2026-07-08T06:17:07.675013407Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:17:22.323874632Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":28,"history_lines":1,"events_offset":264,"events_lines":2,"console_offset":31,"console_lines":1}
+{"time":"2026-07-08T06:17:22.708547057Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:17:37.324045832Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":266,"events_lines":2}
+{"time":"2026-07-08T06:17:37.702534119Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:17:52.324310783Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":268,"events_lines":2}
+{"time":"2026-07-08T06:17:52.69702917Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:18:07.324257367Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":270,"events_lines":2}
+{"time":"2026-07-08T06:18:07.624827715Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:18:22.323701272Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":29,"history_lines":1,"events_offset":272,"events_lines":2,"console_offset":32,"console_lines":1}
+{"time":"2026-07-08T06:18:22.718409029Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:18:37.324231448Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":274,"events_lines":2}
+{"time":"2026-07-08T06:18:37.633438117Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:18:52.323830274Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":276,"events_lines":2}
+{"time":"2026-07-08T06:18:52.619249796Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:19:07.323894835Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":278,"events_lines":2}
+{"time":"2026-07-08T06:19:07.748795737Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:19:22.324025284Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":280,"events_lines":2}
+{"time":"2026-07-08T06:19:22.669690608Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:19:37.323996346Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":30,"history_lines":1,"events_offset":282,"events_lines":2,"console_offset":33,"console_lines":1}
+{"time":"2026-07-08T06:19:37.623048296Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:19:52.323883283Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":284,"events_lines":2}
+{"time":"2026-07-08T06:19:52.628174909Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:20:07.323688997Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":286,"events_lines":2}
+{"time":"2026-07-08T06:20:07.664949411Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:20:22.323464066Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":288,"events_lines":2}
+{"time":"2026-07-08T06:20:22.594954241Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:20:37.32440722Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":31,"history_lines":1,"events_offset":290,"events_lines":2,"console_offset":34,"console_lines":1}
+{"time":"2026-07-08T06:20:37.719879229Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:20:52.324395277Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":292,"events_lines":2}
+{"time":"2026-07-08T06:20:53.117568635Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:21:07.323538251Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":294,"events_lines":2}
+{"time":"2026-07-08T06:21:07.694042892Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:21:22.324385512Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":296,"events_lines":2}
+{"time":"2026-07-08T06:21:22.646422397Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:21:37.324558625Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":32,"history_lines":1,"events_offset":298,"events_lines":2,"console_offset":35,"console_lines":1}
+{"time":"2026-07-08T06:21:37.743009852Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:21:52.323416213Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":300,"events_lines":2}
+{"time":"2026-07-08T06:21:52.654094107Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:22:07.323440905Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":302,"events_lines":2}
+{"time":"2026-07-08T06:22:07.646985352Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:22:22.323604454Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":304,"events_lines":2}
+{"time":"2026-07-08T06:22:22.636794243Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:22:37.324023004Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":33,"history_lines":1,"events_offset":306,"events_lines":2,"console_offset":36,"console_lines":1}
+{"time":"2026-07-08T06:22:37.670184639Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:22:52.323790391Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":308,"events_lines":2}
+{"time":"2026-07-08T06:22:52.728288614Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:23:07.323747962Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":310,"events_lines":2}
+{"time":"2026-07-08T06:23:07.552477738Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}

epfl_final/SparseAtt/wandb/run-20260708_054405-43sa4ouk/logs/debug.log

@@ -0,0 +1,19 @@
+2026-07-08 05:44:05,590 INFO    MainThread:275 [wandb_setup.py:_flush():81] Current SDK version is 0.26.1
+2026-07-08 05:44:05,590 INFO    MainThread:275 [wandb_setup.py:_flush():81] Configure stats pid to 275
+2026-07-08 05:44:05,590 INFO    MainThread:275 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2026-07-08 05:44:05,590 INFO    MainThread:275 [wandb_init.py:setup_run_log_directory():723] Logging user logs to /mloscratch/homes/navasard/sparse/SparseAtt/wandb/run-20260708_054405-43sa4ouk/logs/debug.log
+2026-07-08 05:44:05,590 INFO    MainThread:275 [wandb_init.py:setup_run_log_directory():724] Logging internal logs to /mloscratch/homes/navasard/sparse/SparseAtt/wandb/run-20260708_054405-43sa4ouk/logs/debug-internal.log
+2026-07-08 05:44:05,590 INFO    MainThread:275 [wandb_init.py:init():850] calling init triggers
+2026-07-08 05:44:05,590 INFO    MainThread:275 [wandb_init.py:init():855] wandb.init called with sweep_config: {}
+config: {'model_path': '/mloscratch/homes/navasard/qwen3_06b_base', 'model_dtype': 'bfloat16', 'init_from_scratch': True, 'train_path': '/mloscratch/homes/navasard/data/longctx30k/tokenized', 'val_path': None, 'max_seq_length': 30720, 'num_workers': 4, 'output_dir': '/mloscratch/homes/navasard/sparse/checkpoints/l30k1_2b_dense/run_20260708_054311', 'resume_from_checkpoint': None, 'force_new_run': False, 'num_epochs': 1, 'max_steps': None, 'per_device_batch_size': 1, 'gradient_accumulation_steps': 5, 'learning_rate': 0.0005, 'weight_decay': 0.01, 'warmup_steps': 750, 'lr_scheduler': 'cosine', 'max_grad_norm': 1.0, 'save_steps': 1000, 'eval_steps': 500, 'eval_batches': 20, 'logging_steps': 10, 'seed': 42, 'verbose': True, 'checkpoint_activations': True, 'strategy': 'ddp', 'mixed_precision': 'bf16', 'wandb_enabled': True, 'wandb_project': 'stochastic_sparsity', 'wandb_name': '1.2b dense 30k longctx', 'wandb_tags': ['qwen3', '1.2b', 'dense', '30k', 'longctx', 'b128', 'scaleladder'], 'hidden_size': 1536, 'intermediate_size': 6144, 'num_attention_heads': 12, 'num_key_value_heads': 4, 'num_hidden_layers': None, 'head_dim': None, 'attention_type': 'sparse_stochastic', 'sparse_block_size': 128, 'sparse_num_random_blocks': 120, 'sparse_mask_pool_size': 1, 'sparse_val_dense_compare': False, 'sparse_layer_indices': None, 'sparse_mask_cache_dir': None, 'sparse_mask_seed': 0, 'dense_mode': True, 'dense_step_intervals': None, '_wandb': {}}
+2026-07-08 05:44:05,590 INFO    MainThread:275 [wandb_init.py:init():898] starting backend
+2026-07-08 05:44:06,127 INFO    MainThread:275 [wandb_init.py:init():913] sending inform_init request
+2026-07-08 05:44:06,650 INFO    MainThread:275 [wandb_init.py:init():918] backend started and connected
+2026-07-08 05:44:06,655 INFO    MainThread:275 [wandb_init.py:init():988] updated telemetry
+2026-07-08 05:44:06,682 INFO    MainThread:275 [wandb_init.py:init():1011] communicating run to backend with 90.0 second timeout
+2026-07-08 05:44:07,110 INFO    MainThread:275 [wandb_init.py:init():1056] starting run threads in backend
+2026-07-08 05:44:07,310 INFO    MainThread:275 [wandb_run.py:_console_start():2554] atexit reg
+2026-07-08 05:44:07,310 INFO    MainThread:275 [wandb_run.py:_redirect():2403] redirect: wrap_raw
+2026-07-08 05:44:07,310 INFO    MainThread:275 [wandb_run.py:_redirect():2472] Wrapping output streams.
+2026-07-08 05:44:07,310 INFO    MainThread:275 [wandb_run.py:_redirect():2495] Redirects installed.
+2026-07-08 05:44:07,318 INFO    MainThread:275 [wandb_init.py:init():1094] run started, returning control to user process

epfl_final/SparseAtt/wandb/run-20260708_054405-43sa4ouk/run-43sa4ouk.wandb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ba89a10ccb8462aa1b922476e5219b6fba33d0e2fb65d5b312aeb56d18a19cf2
+size 720896

epfl_final/SparseAtt/wandb/run-20260708_060003-qlqhc3gy/files/config.yaml

@@ -0,0 +1,173 @@
+_wandb:
+    value:
+        cli_version: 0.26.1
+        e:
+            ojn4wwkj2s355fjkagcq96p92b1753oj:
+                args:
+                    - --config
+                    - configs/ladder8k/dense.json
+                codePath: train.py
+                codePathLocal: train.py
+                cpu_count: 96
+                cpu_count_logical: 192
+                cudaVersion: "13.0"
+                disk:
+                    /:
+                        total: "9374931091456"
+                        used: "3858119897088"
+                email: ruben.navasardyan@mail.utoronto.ca
+                executable: /mloscratch/homes/navasard/conda/bin/python
+                git:
+                    commit: 88298c7a418622530ca5426b22b1e5ebd10ca0cb
+                    remote: https://github.com/rubentium/TokenPrune.git
+                gpu: NVIDIA H100 80GB HBM3
+                gpu_count: 1
+                gpu_nvidia:
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-3f5bace4-0066-71eb-863d-73a3a6d79e3b
+                host: h100-workspace-1-gpu-0-0
+                memory:
+                    total: "1623005732864"
+                os: Linux-6.8.0-124-generic-x86_64-with-glibc2.35
+                program: /mloscratch/homes/navasard/sparse/SparseAtt/train.py
+                python: CPython 3.13.13
+                root: /mloscratch/homes/navasard/sparse/SparseAtt
+                startedAt: "2026-07-08T06:00:03.268761Z"
+                writerId: ojn4wwkj2s355fjkagcq96p92b1753oj
+        m: []
+        python_version: 3.13.13
+        t:
+            "1":
+                - 1
+                - 5
+                - 11
+                - 49
+                - 51
+                - 53
+                - 71
+            "2":
+                - 1
+                - 5
+                - 11
+                - 49
+                - 51
+                - 53
+                - 71
+            "3":
+                - 2
+                - 5
+                - 13
+                - 14
+                - 15
+                - 16
+                - 61
+                - 62
+            "4": 3.13.13
+            "5": 0.26.1
+            "6": 5.7.0
+            "12": 0.26.1
+            "13": linux-x86_64
+attention_type:
+    value: sparse_stochastic
+checkpoint_activations:
+    value: false
+dense_mode:
+    value: true
+dense_step_intervals:
+    value: null
+eval_batches:
+    value: 50
+eval_steps:
+    value: 250
+force_new_run:
+    value: false
+gradient_accumulation_steps:
+    value: 10
+head_dim:
+    value: null
+hidden_size:
+    value: 768
+init_from_scratch:
+    value: true
+intermediate_size:
+    value: 3072
+learning_rate:
+    value: 0.0005
+logging_steps:
+    value: 10
+lr_scheduler:
+    value: cosine
+max_grad_norm:
+    value: 1
+max_seq_length:
+    value: 8192
+max_steps:
+    value: 44000
+mixed_precision:
+    value: bf16
+model_dtype:
+    value: bfloat16
+model_path:
+    value: /mloscratch/homes/navasard/qwen3_06b_base
+num_attention_heads:
+    value: 6
+num_epochs:
+    value: 1
+num_hidden_layers:
+    value: 28
+num_key_value_heads:
+    value: 2
+num_workers:
+    value: 8
+output_dir:
+    value: /mloscratch/homes/navasard/sparse/checkpoints/l8k365m_dense/run_20260707_034553
+per_device_batch_size:
+    value: 2
+resume_from_checkpoint:
+    value: /mloscratch/homes/navasard/sparse/checkpoints/l8k365m_dense/run_20260707_034553/checkpoint-42000
+save_steps:
+    value: 1000
+seed:
+    value: 42
+sparse_block_size:
+    value: 128
+sparse_layer_indices:
+    value: null
+sparse_mask_cache_dir:
+    value: null
+sparse_mask_pool_size:
+    value: 1
+sparse_mask_seed:
+    value: 0
+sparse_num_random_blocks:
+    value: 31
+sparse_val_dense_compare:
+    value: false
+strategy:
+    value: ddp
+train_path:
+    value: /mloscratch/homes/navasard/data/fineweb_large/tokenized
+val_path:
+    value: null
+verbose:
+    value: true
+wandb_enabled:
+    value: true
+wandb_name:
+    value: 365m dense 8k
+wandb_project:
+    value: stochastic_sparsity
+wandb_tags:
+    value:
+        - qwen3
+        - 365m
+        - 8k
+        - scaleladder
+        - dense
+warmup_steps:
+    value: 750
+weight_decay:
+    value: 0.01

epfl_final/SparseAtt/wandb/run-20260708_060003-qlqhc3gy/files/output.log

@@ -0,0 +1,304 @@
+Starting training...
+Validation at step 42000: Loss = 2.8956, Perplexity = 18.09
+wandb: WARNING Tried to log to step 42000 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Switching attention mode for step 42001: dense (dense step interval)
+Step 42010/44000 | Loss: 2.9026 | LR: 2.61e-06 | 58442.7ms/step | 0.3 samples/s
+wandb: WARNING Tried to log to step 42010 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42020/44000 | Loss: 2.8940 | LR: 2.58e-06 | 2471.8ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42020 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42030/44000 | Loss: 2.8854 | LR: 2.56e-06 | 2472.3ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42030 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42040/44000 | Loss: 2.8688 | LR: 2.53e-06 | 2473.4ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42040 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42050/44000 | Loss: 2.8986 | LR: 2.50e-06 | 2475.4ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42050 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42060/44000 | Loss: 2.8971 | LR: 2.48e-06 | 2474.1ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42060 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42070/44000 | Loss: 2.9161 | LR: 2.45e-06 | 2473.5ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42070 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42080/44000 | Loss: 2.8754 | LR: 2.43e-06 | 2474.6ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42080 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42090/44000 | Loss: 2.8948 | LR: 2.40e-06 | 2475.4ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42090 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42100/44000 | Loss: 2.8972 | LR: 2.38e-06 | 2476.0ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42100 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42110/44000 | Loss: 2.9053 | LR: 2.35e-06 | 2475.6ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42110 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42120/44000 | Loss: 2.9192 | LR: 2.33e-06 | 2474.4ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42120 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42130/44000 | Loss: 2.8948 | LR: 2.30e-06 | 2473.3ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42130 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42140/44000 | Loss: 2.9002 | LR: 2.28e-06 | 2473.8ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42140 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42150/44000 | Loss: 2.8784 | LR: 2.25e-06 | 2474.0ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42150 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42160/44000 | Loss: 2.9180 | LR: 2.23e-06 | 2473.2ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42160 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42170/44000 | Loss: 2.9129 | LR: 2.21e-06 | 2473.5ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42170 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42180/44000 | Loss: 2.9178 | LR: 2.18e-06 | 2473.3ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42180 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42190/44000 | Loss: 2.9356 | LR: 2.16e-06 | 2473.3ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42190 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42200/44000 | Loss: 2.9144 | LR: 2.13e-06 | 2473.3ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42200 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42210/44000 | Loss: 2.8649 | LR: 2.11e-06 | 2474.5ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42210 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42220/44000 | Loss: 2.9082 | LR: 2.09e-06 | 2475.2ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42220 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42230/44000 | Loss: 2.9055 | LR: 2.06e-06 | 2474.6ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42230 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42240/44000 | Loss: 2.8562 | LR: 2.04e-06 | 2474.7ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42240 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42250/44000 | Loss: 2.8947 | LR: 2.02e-06 | 2474.9ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42250 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Validation at step 42250: Loss = 2.9152, Perplexity = 18.45
+Step 42260/44000 | Loss: 2.9025 | LR: 1.99e-06 | 6679.7ms/step | 3.0 samples/s
+wandb: WARNING Tried to log to step 42260 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42270/44000 | Loss: 2.9179 | LR: 1.97e-06 | 2474.1ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42270 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42280/44000 | Loss: 2.8910 | LR: 1.95e-06 | 2474.6ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42280 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42290/44000 | Loss: 2.9246 | LR: 1.93e-06 | 2474.4ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42290 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42300/44000 | Loss: 2.8877 | LR: 1.90e-06 | 2473.5ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42300 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42310/44000 | Loss: 2.9260 | LR: 1.88e-06 | 2474.1ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42310 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42320/44000 | Loss: 2.9189 | LR: 1.86e-06 | 2473.8ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42320 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42330/44000 | Loss: 2.9417 | LR: 1.84e-06 | 2474.1ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42330 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42340/44000 | Loss: 2.8960 | LR: 1.82e-06 | 2474.1ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42340 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42350/44000 | Loss: 2.8536 | LR: 1.79e-06 | 2474.2ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42350 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42360/44000 | Loss: 2.9162 | LR: 1.77e-06 | 2474.4ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42360 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42370/44000 | Loss: 2.9019 | LR: 1.75e-06 | 2473.8ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42370 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42380/44000 | Loss: 2.9031 | LR: 1.73e-06 | 2473.9ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42380 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42390/44000 | Loss: 2.9117 | LR: 1.71e-06 | 2472.9ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42390 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42400/44000 | Loss: 2.8752 | LR: 1.69e-06 | 2472.6ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42400 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42410/44000 | Loss: 2.8923 | LR: 1.67e-06 | 2477.5ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42410 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42420/44000 | Loss: 2.8904 | LR: 1.64e-06 | 2478.4ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42420 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42430/44000 | Loss: 2.8909 | LR: 1.62e-06 | 2474.9ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42430 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42440/44000 | Loss: 2.8826 | LR: 1.60e-06 | 2473.5ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42440 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42450/44000 | Loss: 2.8496 | LR: 1.58e-06 | 2472.3ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42450 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42460/44000 | Loss: 2.8867 | LR: 1.56e-06 | 2473.0ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42460 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42470/44000 | Loss: 2.8888 | LR: 1.54e-06 | 2473.8ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42470 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42480/44000 | Loss: 2.8738 | LR: 1.52e-06 | 2474.0ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42480 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42490/44000 | Loss: 2.9266 | LR: 1.50e-06 | 2474.3ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42490 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42500/44000 | Loss: 2.9400 | LR: 1.48e-06 | 2473.3ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42500 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Validation at step 42500: Loss = 2.9134, Perplexity = 18.42
+Step 42510/44000 | Loss: 2.8661 | LR: 1.46e-06 | 6676.2ms/step | 3.0 samples/s
+wandb: WARNING Tried to log to step 42510 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42520/44000 | Loss: 2.8836 | LR: 1.44e-06 | 2473.9ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42520 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42530/44000 | Loss: 2.8974 | LR: 1.42e-06 | 2473.8ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42530 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42540/44000 | Loss: 2.8869 | LR: 1.40e-06 | 2473.9ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42540 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42550/44000 | Loss: 2.9057 | LR: 1.39e-06 | 2473.6ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42550 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42560/44000 | Loss: 2.9137 | LR: 1.37e-06 | 2476.8ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42560 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42570/44000 | Loss: 2.9186 | LR: 1.35e-06 | 2474.4ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42570 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42580/44000 | Loss: 2.9066 | LR: 1.33e-06 | 2472.8ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42580 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42590/44000 | Loss: 2.9172 | LR: 1.31e-06 | 2473.4ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42590 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42600/44000 | Loss: 2.9135 | LR: 1.29e-06 | 2472.6ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42600 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42610/44000 | Loss: 2.8956 | LR: 1.27e-06 | 2474.9ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42610 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42620/44000 | Loss: 2.8944 | LR: 1.25e-06 | 2475.7ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42620 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42630/44000 | Loss: 2.8900 | LR: 1.24e-06 | 2474.4ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42630 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42640/44000 | Loss: 2.9035 | LR: 1.22e-06 | 2474.4ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42640 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42650/44000 | Loss: 2.8893 | LR: 1.20e-06 | 2477.1ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42650 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42660/44000 | Loss: 2.9111 | LR: 1.18e-06 | 2474.4ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42660 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42670/44000 | Loss: 2.8768 | LR: 1.17e-06 | 2473.4ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42670 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42680/44000 | Loss: 2.8596 | LR: 1.15e-06 | 2474.7ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42680 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42690/44000 | Loss: 2.8910 | LR: 1.13e-06 | 2475.0ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42690 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42700/44000 | Loss: 2.9024 | LR: 1.11e-06 | 2474.9ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42700 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42710/44000 | Loss: 2.9106 | LR: 1.10e-06 | 2475.7ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42710 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42720/44000 | Loss: 2.9081 | LR: 1.08e-06 | 2475.6ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42720 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42730/44000 | Loss: 2.8979 | LR: 1.06e-06 | 2475.9ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42730 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42740/44000 | Loss: 2.9098 | LR: 1.05e-06 | 2473.4ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42740 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42750/44000 | Loss: 2.9176 | LR: 1.03e-06 | 2474.7ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42750 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Validation at step 42750: Loss = 2.9128, Perplexity = 18.41
+Step 42760/44000 | Loss: 2.8979 | LR: 1.01e-06 | 6675.5ms/step | 3.0 samples/s
+wandb: WARNING Tried to log to step 42760 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42770/44000 | Loss: 2.9136 | LR: 9.97e-07 | 2473.9ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42770 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42780/44000 | Loss: 2.8889 | LR: 9.81e-07 | 2474.6ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42780 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42790/44000 | Loss: 2.9061 | LR: 9.65e-07 | 2475.3ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42790 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42800/44000 | Loss: 2.9060 | LR: 9.49e-07 | 2474.3ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42800 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42810/44000 | Loss: 2.8747 | LR: 9.33e-07 | 2474.0ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42810 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42820/44000 | Loss: 2.9063 | LR: 9.18e-07 | 2474.3ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42820 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42830/44000 | Loss: 2.9125 | LR: 9.02e-07 | 2475.6ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42830 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42840/44000 | Loss: 2.9042 | LR: 8.87e-07 | 2474.3ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42840 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42850/44000 | Loss: 2.8695 | LR: 8.72e-07 | 2475.6ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42850 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42860/44000 | Loss: 2.8815 | LR: 8.57e-07 | 2478.8ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42860 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42870/44000 | Loss: 2.9173 | LR: 8.42e-07 | 2474.9ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42870 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42880/44000 | Loss: 2.9004 | LR: 8.27e-07 | 2472.8ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42880 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42890/44000 | Loss: 2.9127 | LR: 8.12e-07 | 2472.7ms/step | 8.1 samples/s
+wandb: WARNING Tried to log to step 42890 that is less than the current step 42891. Steps must be monotonically increasing, so this data will be ignored. See https://wandb.me/define-metric to log data out of order.
+Step 42900/44000 | Loss: 2.9098 | LR: 7.98e-07 | 2472.7ms/step | 8.1 samples/s
+Step 42910/44000 | Loss: 2.8639 | LR: 7.83e-07 | 2472.6ms/step | 8.1 samples/s
+Step 42920/44000 | Loss: 2.8815 | LR: 7.69e-07 | 2473.5ms/step | 8.1 samples/s
+Step 42930/44000 | Loss: 2.8930 | LR: 7.55e-07 | 2473.4ms/step | 8.1 samples/s
+Step 42940/44000 | Loss: 2.8811 | LR: 7.41e-07 | 2473.3ms/step | 8.1 samples/s
+Step 42950/44000 | Loss: 2.9038 | LR: 7.27e-07 | 2473.6ms/step | 8.1 samples/s
+Step 42960/44000 | Loss: 2.8809 | LR: 7.13e-07 | 2472.9ms/step | 8.1 samples/s
+Step 42970/44000 | Loss: 2.9096 | LR: 6.99e-07 | 2472.7ms/step | 8.1 samples/s
+Step 42980/44000 | Loss: 2.9116 | LR: 6.86e-07 | 2472.9ms/step | 8.1 samples/s
+Step 42990/44000 | Loss: 2.8679 | LR: 6.72e-07 | 2473.2ms/step | 8.1 samples/s
+Step 43000/44000 | Loss: 2.8914 | LR: 6.59e-07 | 2473.3ms/step | 8.1 samples/s
+Validation at step 43000: Loss = 2.9128, Perplexity = 18.41
+Saved checkpoint to /mloscratch/homes/navasard/sparse/checkpoints/l8k365m_dense/run_20260707_034553/checkpoint-43000
+Step 43010/44000 | Loss: 2.8783 | LR: 6.46e-07 | 6981.5ms/step | 2.9 samples/s
+Step 43020/44000 | Loss: 2.9055 | LR: 6.33e-07 | 2473.4ms/step | 8.1 samples/s
+Step 43030/44000 | Loss: 2.9116 | LR: 6.20e-07 | 2473.2ms/step | 8.1 samples/s
+Step 43040/44000 | Loss: 2.8814 | LR: 6.08e-07 | 2472.8ms/step | 8.1 samples/s
+Step 43050/44000 | Loss: 2.8901 | LR: 5.95e-07 | 2473.4ms/step | 8.1 samples/s
+Step 43060/44000 | Loss: 2.8589 | LR: 5.83e-07 | 2473.9ms/step | 8.1 samples/s
+Step 43070/44000 | Loss: 2.8816 | LR: 5.70e-07 | 2474.0ms/step | 8.1 samples/s
+Step 43080/44000 | Loss: 2.8898 | LR: 5.58e-07 | 2474.0ms/step | 8.1 samples/s
+Step 43090/44000 | Loss: 2.9181 | LR: 5.46e-07 | 2474.5ms/step | 8.1 samples/s
+Step 43100/44000 | Loss: 2.8988 | LR: 5.34e-07 | 2473.4ms/step | 8.1 samples/s
+Step 43110/44000 | Loss: 2.8700 | LR: 5.22e-07 | 2472.6ms/step | 8.1 samples/s
+Step 43120/44000 | Loss: 2.8807 | LR: 5.11e-07 | 2472.0ms/step | 8.1 samples/s
+Step 43130/44000 | Loss: 2.9162 | LR: 4.99e-07 | 2473.0ms/step | 8.1 samples/s
+Step 43140/44000 | Loss: 2.9092 | LR: 4.88e-07 | 2472.3ms/step | 8.1 samples/s
+Step 43150/44000 | Loss: 2.8604 | LR: 4.76e-07 | 2472.6ms/step | 8.1 samples/s
+Step 43160/44000 | Loss: 2.8885 | LR: 4.65e-07 | 2471.5ms/step | 8.1 samples/s
+Step 43170/44000 | Loss: 2.8890 | LR: 4.54e-07 | 2472.7ms/step | 8.1 samples/s
+Step 43180/44000 | Loss: 2.8879 | LR: 4.43e-07 | 2474.7ms/step | 8.1 samples/s
+Step 43190/44000 | Loss: 2.8875 | LR: 4.33e-07 | 2473.3ms/step | 8.1 samples/s
+Step 43200/44000 | Loss: 2.9041 | LR: 4.22e-07 | 2473.3ms/step | 8.1 samples/s
+Step 43210/44000 | Loss: 2.8635 | LR: 4.12e-07 | 2473.5ms/step | 8.1 samples/s
+Step 43220/44000 | Loss: 2.9215 | LR: 4.01e-07 | 2473.8ms/step | 8.1 samples/s
+Step 43230/44000 | Loss: 2.9011 | LR: 3.91e-07 | 2473.3ms/step | 8.1 samples/s
+Step 43240/44000 | Loss: 2.9095 | LR: 3.81e-07 | 2473.7ms/step | 8.1 samples/s
+Step 43250/44000 | Loss: 2.9180 | LR: 3.71e-07 | 2473.4ms/step | 8.1 samples/s
+Validation at step 43250: Loss = 2.9128, Perplexity = 18.41
+Step 43260/44000 | Loss: 2.9016 | LR: 3.61e-07 | 6676.0ms/step | 3.0 samples/s
+Step 43270/44000 | Loss: 2.9167 | LR: 3.51e-07 | 2472.6ms/step | 8.1 samples/s
+Step 43280/44000 | Loss: 2.8757 | LR: 3.42e-07 | 2472.6ms/step | 8.1 samples/s
+Step 43290/44000 | Loss: 2.8804 | LR: 3.32e-07 | 2472.1ms/step | 8.1 samples/s
+Step 43300/44000 | Loss: 2.9320 | LR: 3.23e-07 | 2471.2ms/step | 8.1 samples/s
+Step 43310/44000 | Loss: 2.8923 | LR: 3.14e-07 | 2471.9ms/step | 8.1 samples/s
+Step 43320/44000 | Loss: 2.8882 | LR: 3.05e-07 | 2480.6ms/step | 8.1 samples/s
+Step 43330/44000 | Loss: 2.9103 | LR: 2.96e-07 | 2472.7ms/step | 8.1 samples/s
+Step 43340/44000 | Loss: 2.8948 | LR: 2.87e-07 | 2473.0ms/step | 8.1 samples/s
+Step 43350/44000 | Loss: 2.8576 | LR: 2.79e-07 | 2473.1ms/step | 8.1 samples/s
+Step 43360/44000 | Loss: 2.8828 | LR: 2.70e-07 | 2473.6ms/step | 8.1 samples/s
+Step 43370/44000 | Loss: 2.8778 | LR: 2.62e-07 | 2474.4ms/step | 8.1 samples/s
+Step 43380/44000 | Loss: 2.8832 | LR: 2.53e-07 | 2474.2ms/step | 8.1 samples/s
+Step 43390/44000 | Loss: 2.8876 | LR: 2.45e-07 | 2473.4ms/step | 8.1 samples/s
+Step 43400/44000 | Loss: 2.8796 | LR: 2.37e-07 | 2475.7ms/step | 8.1 samples/s
+Step 43410/44000 | Loss: 2.8704 | LR: 2.30e-07 | 2480.9ms/step | 8.1 samples/s
+Step 43420/44000 | Loss: 2.9239 | LR: 2.22e-07 | 2479.2ms/step | 8.1 samples/s
+Step 43430/44000 | Loss: 2.8647 | LR: 2.14e-07 | 2482.4ms/step | 8.1 samples/s
+Step 43440/44000 | Loss: 2.8877 | LR: 2.07e-07 | 2481.7ms/step | 8.1 samples/s
+Step 43450/44000 | Loss: 2.8725 | LR: 1.99e-07 | 2479.4ms/step | 8.1 samples/s
+Step 43460/44000 | Loss: 2.8947 | LR: 1.92e-07 | 2475.6ms/step | 8.1 samples/s
+Step 43470/44000 | Loss: 2.8734 | LR: 1.85e-07 | 2477.6ms/step | 8.1 samples/s
+Step 43480/44000 | Loss: 2.8769 | LR: 1.78e-07 | 2477.4ms/step | 8.1 samples/s
+Step 43490/44000 | Loss: 2.9095 | LR: 1.72e-07 | 2476.5ms/step | 8.1 samples/s
+Step 43500/44000 | Loss: 2.8874 | LR: 1.65e-07 | 2473.5ms/step | 8.1 samples/s
+Validation at step 43500: Loss = 2.9128, Perplexity = 18.41
+Step 43510/44000 | Loss: 2.8898 | LR: 1.58e-07 | 6675.9ms/step | 3.0 samples/s
+Step 43520/44000 | Loss: 2.8896 | LR: 1.52e-07 | 2471.9ms/step | 8.1 samples/s
+Step 43530/44000 | Loss: 2.8807 | LR: 1.46e-07 | 2472.7ms/step | 8.1 samples/s
+Step 43540/44000 | Loss: 2.8856 | LR: 1.40e-07 | 2472.7ms/step | 8.1 samples/s
+Step 43550/44000 | Loss: 2.9021 | LR: 1.34e-07 | 2472.8ms/step | 8.1 samples/s
+Step 43560/44000 | Loss: 2.8786 | LR: 1.28e-07 | 2473.0ms/step | 8.1 samples/s
+Step 43570/44000 | Loss: 2.8951 | LR: 1.22e-07 | 2472.2ms/step | 8.1 samples/s
+Step 43580/44000 | Loss: 2.9028 | LR: 1.16e-07 | 2472.8ms/step | 8.1 samples/s
+Step 43590/44000 | Loss: 2.9086 | LR: 1.11e-07 | 2472.4ms/step | 8.1 samples/s
+Step 43600/44000 | Loss: 2.8982 | LR: 1.06e-07 | 2472.2ms/step | 8.1 samples/s
+Step 43610/44000 | Loss: 2.8895 | LR: 1.00e-07 | 2472.7ms/step | 8.1 samples/s
+Step 43620/44000 | Loss: 2.8601 | LR: 9.52e-08 | 2473.7ms/step | 8.1 samples/s
+Step 43630/44000 | Loss: 2.9137 | LR: 9.03e-08 | 2474.1ms/step | 8.1 samples/s
+Step 43640/44000 | Loss: 2.9201 | LR: 8.55e-08 | 2479.9ms/step | 8.1 samples/s
+Step 43650/44000 | Loss: 2.8716 | LR: 8.08e-08 | 2475.0ms/step | 8.1 samples/s
+Step 43660/44000 | Loss: 2.8757 | LR: 7.62e-08 | 2473.3ms/step | 8.1 samples/s
+Step 43670/44000 | Loss: 2.9088 | LR: 7.18e-08 | 2473.4ms/step | 8.1 samples/s
+Step 43680/44000 | Loss: 2.9063 | LR: 6.75e-08 | 2476.3ms/step | 8.1 samples/s
+Step 43690/44000 | Loss: 2.9049 | LR: 6.34e-08 | 2474.2ms/step | 8.1 samples/s
+Step 43700/44000 | Loss: 2.9062 | LR: 5.94e-08 | 2472.7ms/step | 8.1 samples/s
+Step 43710/44000 | Loss: 2.8842 | LR: 5.55e-08 | 2472.9ms/step | 8.1 samples/s
+Step 43720/44000 | Loss: 2.8903 | LR: 5.17e-08 | 2472.2ms/step | 8.1 samples/s
+Step 43730/44000 | Loss: 2.9026 | LR: 4.81e-08 | 2472.5ms/step | 8.1 samples/s
+Step 43740/44000 | Loss: 2.8897 | LR: 4.46e-08 | 2473.9ms/step | 8.1 samples/s
+Step 43750/44000 | Loss: 2.8944 | LR: 4.12e-08 | 2473.0ms/step | 8.1 samples/s
+Validation at step 43750: Loss = 2.9127, Perplexity = 18.41
+Step 43760/44000 | Loss: 2.8659 | LR: 3.80e-08 | 6671.9ms/step | 3.0 samples/s
+Step 43770/44000 | Loss: 2.8963 | LR: 3.49e-08 | 2475.3ms/step | 8.1 samples/s
+Step 43780/44000 | Loss: 2.8845 | LR: 3.19e-08 | 2475.8ms/step | 8.1 samples/s
+Step 43790/44000 | Loss: 2.9188 | LR: 2.91e-08 | 2477.7ms/step | 8.1 samples/s
+Step 43800/44000 | Loss: 2.8794 | LR: 2.64e-08 | 2474.1ms/step | 8.1 samples/s
+Step 43810/44000 | Loss: 2.8787 | LR: 2.38e-08 | 2474.4ms/step | 8.1 samples/s
+Step 43820/44000 | Loss: 2.8945 | LR: 2.14e-08 | 2474.2ms/step | 8.1 samples/s
+Step 43830/44000 | Loss: 2.8698 | LR: 1.91e-08 | 2474.8ms/step | 8.1 samples/s
+Step 43840/44000 | Loss: 2.8995 | LR: 1.69e-08 | 2474.3ms/step | 8.1 samples/s
+Step 43850/44000 | Loss: 2.8989 | LR: 1.48e-08 | 2474.5ms/step | 8.1 samples/s
+Step 43860/44000 | Loss: 2.9028 | LR: 1.29e-08 | 2474.0ms/step | 8.1 samples/s
+Step 43870/44000 | Loss: 2.8806 | LR: 1.11e-08 | 2473.2ms/step | 8.1 samples/s
+Step 43880/44000 | Loss: 2.8919 | LR: 9.50e-09 | 2475.0ms/step | 8.1 samples/s
+Step 43890/44000 | Loss: 2.9090 | LR: 7.98e-09 | 2473.2ms/step | 8.1 samples/s
+Step 43900/44000 | Loss: 2.8898 | LR: 6.60e-09 | 2475.8ms/step | 8.1 samples/s
+Step 43910/44000 | Loss: 2.9074 | LR: 5.34e-09 | 2474.6ms/step | 8.1 samples/s
+Step 43920/44000 | Loss: 2.8697 | LR: 4.22e-09 | 2476.8ms/step | 8.1 samples/s
+Step 43930/44000 | Loss: 2.9084 | LR: 3.23e-09 | 2474.6ms/step | 8.1 samples/s
+Step 43940/44000 | Loss: 2.8652 | LR: 2.37e-09 | 2478.6ms/step | 8.1 samples/s
+Step 43950/44000 | Loss: 2.9014 | LR: 1.65e-09 | 2473.5ms/step | 8.1 samples/s
+Step 43960/44000 | Loss: 2.8861 | LR: 1.06e-09 | 2473.2ms/step | 8.1 samples/s
+Step 43970/44000 | Loss: 2.9254 | LR: 5.94e-10 | 2474.1ms/step | 8.1 samples/s
+Step 43980/44000 | Loss: 2.8674 | LR: 2.64e-10 | 2473.4ms/step | 8.1 samples/s
+Step 43990/44000 | Loss: 2.9117 | LR: 6.60e-11 | 2474.6ms/step | 8.1 samples/s
+Step 44000/44000 | Loss: 2.8938 | LR: 0.00e+00 | 2472.7ms/step | 8.1 samples/s
+Validation at step 44000: Loss = 2.9128, Perplexity = 18.41
+Saved checkpoint to /mloscratch/homes/navasard/sparse/checkpoints/l8k365m_dense/run_20260707_034553/checkpoint-44000
+Saved checkpoint to /mloscratch/homes/navasard/sparse/checkpoints/l8k365m_dense/run_20260707_034553/checkpoint-44000

epfl_final/SparseAtt/wandb/run-20260708_060003-qlqhc3gy/files/requirements.txt

@@ -0,0 +1,141 @@
+pip==26.0.1
+nvidia-cusparselt-cu13==0.8.0
+mpmath==1.3.0
+cuda-toolkit==13.0.2
+zstandard==0.25.0
+xxhash==3.7.0
+urllib3==2.6.3
+typing_extensions==4.15.0
+triton==3.6.0
+tqdm==4.67.3
+torchinfo==1.8.0
+sympy==1.14.0
+smmap==5.0.3
+six==1.17.0
+shellingham==1.5.4
+setuptools==81.0.0
+safetensors==0.7.0
+regex==2026.4.4
+PyYAML==6.0.3
+Pygments==2.20.0
+pyarrow==24.0.0
+psutil==7.2.2
+protobuf==7.34.1
+propcache==0.4.1
+platformdirs==4.9.6
+packaging==26.2
+nvidia-nvtx==13.0.85
+nvidia-nvshmem-cu13==3.4.5
+nvidia-nvjitlink==13.0.88
+nvidia-nccl-cu13==2.28.9
+nvidia-curand==10.4.0.35
+nvidia-cufile==1.15.1.6
+nvidia-cuda-runtime==13.0.96
+nvidia-cuda-nvrtc==13.0.88
+nvidia-cuda-cupti==13.0.85
+nvidia-cublas==13.1.0.3
+numpy==2.4.4
+networkx==3.6.1
+multidict==6.7.1
+mdurl==0.1.2
+MarkupSafe==3.0.3
+idna==3.13
+hf-xet==1.4.3
+h11==0.16.0
+fsspec==2026.2.0
+frozenlist==1.8.0
+filelock==3.29.0
+dill==0.4.1
+cuda-pathfinder==1.5.4
+click==8.3.3
+charset-normalizer==3.4.7
+certifi==2026.4.22
+attrs==26.1.0
+annotated-types==0.7.0
+annotated-doc==0.0.4
+aiohappyeyeballs==2.6.1
+yarl==1.23.0
+typing-inspection==0.4.2
+sentry-sdk==2.58.0
+requests==2.33.1
+python-dateutil==2.9.0.post0
+pydantic_core==2.46.3
+nvidia-cusparse==12.6.3.3
+nvidia-cufft==12.0.0.61
+nvidia-cudnn-cu13==9.19.0.56
+multiprocess==0.70.19
+markdown-it-py==4.0.0
+Jinja2==3.1.6
+httpcore==1.0.9
+gitdb==4.0.12
+cuda-bindings==13.2.0
+anyio==4.13.0
+aiosignal==1.4.0
+rich==15.0.0
+pydantic==2.13.3
+pandas==3.0.2
+nvidia-cusolver==12.0.4.66
+httpx==0.28.1
+GitPython==3.1.49
+aiohttp==3.13.5
+wandb==0.26.1
+typer==0.25.1
+torch==2.11.0
+huggingface_hub==1.13.0
+tokenizers==0.22.2
+datasets==4.8.5
+accelerate==1.13.0
+transformers==5.7.0
+word2number==1.1
+sqlitedict==2.1.0
+pytz==2026.2
+threadpoolctl==3.6.0
+tcolorpy==0.1.7
+tabulate==0.10.0
+scipy==1.17.1
+portalocker==3.2.0
+pathvalidate==3.3.1
+narwhals==2.22.1
+more-itertools==11.1.0
+lxml==6.1.1
+joblib==1.5.3
+colorama==0.4.6
+chardet==6.0.0.post1
+absl-py==2.4.0
+scikit-learn==1.9.0
+sacrebleu==2.6.0
+nltk==3.9.4
+mbstrdecoder==1.1.5
+typepy==1.3.5
+rouge_score==0.1.2
+DataProperty==1.1.1
+tabledata==1.3.5
+evaluate==0.4.6
+pytablewriter==1.2.1
+lm_eval==0.4.12
+pluggy==1.6.0
+iniconfig==2.3.0
+pytest==9.1.0
+py-spy==0.4.2
+pillow==12.2.0
+tifffile==2026.6.1
+imagecodecs==2026.6.26
+zipp==4.1.0
+importlib_metadata==9.0.0
+derivative==0.6.3
+pysindy==2.1.0
+pyparsing==3.3.2
+kiwisolver==1.5.0
+fonttools==4.63.0
+cycler==0.12.1
+contourpy==1.3.3
+matplotlib==3.11.0
+sortedcontainers==2.4.0
+wcwidth==0.8.2
+sigfig==1.3.19
+importlib_resources==7.1.0
+enum-tools==0.13.0
+cramjam==2.11.0
+python-snappy==0.7.3
+compact-json==1.8.2
+numbers-parser==4.18.5

epfl_final/SparseAtt/wandb/run-20260708_060003-qlqhc3gy/files/wandb-metadata.json

@@ -0,0 +1,44 @@
+{
+  "os":  "Linux-6.8.0-124-generic-x86_64-with-glibc2.35",
+  "python":  "CPython 3.13.13",
+  "startedAt":  "2026-07-08T06:00:03.268761Z",
+  "args":  [
+    "--config",
+    "configs/ladder8k/dense.json"
+  ],
+  "program":  "/mloscratch/homes/navasard/sparse/SparseAtt/train.py",
+  "codePath":  "train.py",
+  "codePathLocal":  "train.py",
+  "git":  {
+    "remote":  "https://github.com/rubentium/TokenPrune.git",
+    "commit":  "88298c7a418622530ca5426b22b1e5ebd10ca0cb"
+  },
+  "email":  "ruben.navasardyan@mail.utoronto.ca",
+  "root":  "/mloscratch/homes/navasard/sparse/SparseAtt",
+  "host":  "h100-workspace-1-gpu-0-0",
+  "executable":  "/mloscratch/homes/navasard/conda/bin/python",
+  "cpu_count":  96,
+  "cpu_count_logical":  192,
+  "gpu":  "NVIDIA H100 80GB HBM3",
+  "gpu_count":  1,
+  "disk":  {
+    "/":  {
+      "total":  "9374931091456",
+      "used":  "3858119897088"
+    }
+  },
+  "memory":  {
+    "total":  "1623005732864"
+  },
+  "gpu_nvidia":  [
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-3f5bace4-0066-71eb-863d-73a3a6d79e3b"
+    }
+  ],
+  "cudaVersion":  "13.0",
+  "writerId":  "ojn4wwkj2s355fjkagcq96p92b1753oj"
+}

epfl_final/SparseAtt/wandb/run-20260708_060003-qlqhc3gy/files/wandb-summary.json

@@ -0,0 +1 @@
+{"_timestamp":1.783496252332237e+09,"eval/loss":2.912755045413971,"train/loss":2.8938342690467835,"train/learning_rate":0,"eval/uses_sparse_attention":0,"_step":44000,"_wandb":{"runtime":72222},"eval/dense_interval_start":-1,"train/epoch":0.2906439083679023,"speed/samples_per_sec":8.088320994748951,"eval/dense_interval_end":-1,"train/global_step":44000,"speed/steps_per_sec":0.40441604973744755,"speed/ms_per_step":2472.701072692871,"eval/uses_dense_attention":1,"_runtime":72222.861612867,"eval/dense_interval_active":0,"eval/perplexity":18.40744211132422}

epfl_final/SparseAtt/wandb/run-20260708_060003-qlqhc3gy/logs/debug.log

@@ -0,0 +1,25 @@
+2026-07-08 06:00:03,289 INFO    MainThread:1916 [wandb_setup.py:_flush():81] Current SDK version is 0.26.1
+2026-07-08 06:00:03,289 INFO    MainThread:1916 [wandb_setup.py:_flush():81] Configure stats pid to 1916
+2026-07-08 06:00:03,289 INFO    MainThread:1916 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2026-07-08 06:00:03,289 INFO    MainThread:1916 [wandb_init.py:setup_run_log_directory():723] Logging user logs to /mloscratch/homes/navasard/sparse/SparseAtt/wandb/run-20260708_060003-qlqhc3gy/logs/debug.log
+2026-07-08 06:00:03,289 INFO    MainThread:1916 [wandb_init.py:setup_run_log_directory():724] Logging internal logs to /mloscratch/homes/navasard/sparse/SparseAtt/wandb/run-20260708_060003-qlqhc3gy/logs/debug-internal.log
+2026-07-08 06:00:03,289 INFO    MainThread:1916 [wandb_init.py:init():850] calling init triggers
+2026-07-08 06:00:03,290 INFO    MainThread:1916 [wandb_init.py:init():855] wandb.init called with sweep_config: {}
+config: {'model_path': '/mloscratch/homes/navasard/qwen3_06b_base', 'model_dtype': 'bfloat16', 'init_from_scratch': True, 'train_path': '/mloscratch/homes/navasard/data/fineweb_large/tokenized', 'val_path': None, 'max_seq_length': 8192, 'num_workers': 8, 'output_dir': '/mloscratch/homes/navasard/sparse/checkpoints/l8k365m_dense/run_20260707_034553', 'resume_from_checkpoint': '/mloscratch/homes/navasard/sparse/checkpoints/l8k365m_dense/run_20260707_034553/checkpoint-42000', 'force_new_run': False, 'num_epochs': 1, 'max_steps': 44000, 'per_device_batch_size': 2, 'gradient_accumulation_steps': 10, 'learning_rate': 0.0005, 'weight_decay': 0.01, 'warmup_steps': 750, 'lr_scheduler': 'cosine', 'max_grad_norm': 1.0, 'save_steps': 1000, 'eval_steps': 250, 'eval_batches': 50, 'logging_steps': 10, 'seed': 42, 'verbose': True, 'checkpoint_activations': False, 'strategy': 'ddp', 'mixed_precision': 'bf16', 'wandb_enabled': True, 'wandb_project': 'stochastic_sparsity', 'wandb_name': '365m dense 8k', 'wandb_tags': ['qwen3', '365m', '8k', 'scaleladder', 'dense'], 'hidden_size': 768, 'intermediate_size': 3072, 'num_attention_heads': 6, 'num_key_value_heads': 2, 'num_hidden_layers': 28, 'head_dim': None, 'attention_type': 'sparse_stochastic', 'sparse_block_size': 128, 'sparse_num_random_blocks': 31, 'sparse_mask_pool_size': 1, 'sparse_val_dense_compare': False, 'sparse_layer_indices': None, 'sparse_mask_cache_dir': None, 'sparse_mask_seed': 0, 'dense_mode': True, 'dense_step_intervals': None, '_wandb': {}}
+2026-07-08 06:00:03,290 INFO    MainThread:1916 [wandb_init.py:init():898] starting backend
+2026-07-08 06:00:03,548 INFO    MainThread:1916 [wandb_init.py:init():913] sending inform_init request
+2026-07-08 06:00:04,103 INFO    MainThread:1916 [wandb_init.py:init():918] backend started and connected
+2026-07-08 06:00:04,108 INFO    MainThread:1916 [wandb_init.py:init():988] updated telemetry
+2026-07-08 06:00:04,123 INFO    MainThread:1916 [wandb_init.py:init():1011] communicating run to backend with 90.0 second timeout
+2026-07-08 06:00:04,790 INFO    MainThread:1916 [wandb_init.py:init():1051] run resumed
+2026-07-08 06:00:04,792 INFO    MainThread:1916 [wandb_init.py:init():1056] starting run threads in backend
+2026-07-08 06:00:05,001 INFO    MainThread:1916 [wandb_run.py:_console_start():2554] atexit reg
+2026-07-08 06:00:05,002 INFO    MainThread:1916 [wandb_run.py:_redirect():2403] redirect: wrap_raw
+2026-07-08 06:00:05,002 INFO    MainThread:1916 [wandb_run.py:_redirect():2472] Wrapping output streams.
+2026-07-08 06:00:05,002 INFO    MainThread:1916 [wandb_run.py:_redirect():2495] Redirects installed.
+2026-07-08 06:00:05,009 INFO    MainThread:1916 [wandb_init.py:init():1094] run started, returning control to user process
+2026-07-08 07:37:38,652 INFO    MainThread:1916 [wandb_run.py:_finish():2321] finishing run ruben-navasardyan4869-epfl/stochastic_sparsity/qlqhc3gy
+2026-07-08 07:37:38,653 INFO    MainThread:1916 [wandb_run.py:_atexit_cleanup():2520] got exitcode: 0
+2026-07-08 07:37:38,654 INFO    MainThread:1916 [wandb_run.py:_restore():2502] restore
+2026-07-08 07:37:38,654 INFO    MainThread:1916 [wandb_run.py:_restore():2508] restore done
+2026-07-08 07:37:40,253 INFO    MainThread:1916 [wandb_run.py:_footer_sync_info():3889] logging synced files

epfl_final/SparseAtt/wandb/run-20260708_060003-qlqhc3gy/run-qlqhc3gy.wandb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:83eb3b383a5aed475aad30a5a330ae7d84b3f1e0c6b493a15a6360e67db41957
+size 896388

epfl_final/SparseAtt/wandb/run-20260708_063331-i61v7k8r/files/wandb-metadata.json

@@ -0,0 +1,65 @@
+{
+  "os":  "Linux-6.8.0-124-generic-x86_64-with-glibc2.35",
+  "python":  "CPython 3.13.13",
+  "startedAt":  "2026-07-08T06:33:31.022887Z",
+  "args":  [
+    "--config",
+    "configs/train_config_30k_dense_b128.json"
+  ],
+  "program":  "/mloscratch/homes/navasard/sparse/SparseAtt/train.py",
+  "codePath":  "train.py",
+  "codePathLocal":  "train.py",
+  "git":  {
+    "remote":  "https://github.com/rubentium/TokenPrune.git",
+    "commit":  "88298c7a418622530ca5426b22b1e5ebd10ca0cb"
+  },
+  "email":  "ruben.navasardyan@mail.utoronto.ca",
+  "root":  "/mloscratch/homes/navasard/sparse/SparseAtt",
+  "host":  "h100-train-4-gpu-2-0-2",
+  "executable":  "/mloscratch/homes/navasard/conda/bin/python3.13",
+  "cpu_count":  96,
+  "cpu_count_logical":  192,
+  "gpu":  "NVIDIA H100 80GB HBM3",
+  "gpu_count":  4,
+  "disk":  {
+    "/":  {
+      "total":  "9374931091456",
+      "used":  "5566603927552"
+    }
+  },
+  "memory":  {
+    "total":  "1623005732864"
+  },
+  "gpu_nvidia":  [
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-51f13867-a72c-7b65-ff60-661787f0c18a"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-bab1e315-c967-43cf-bdfb-cac1a8dfdfac"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-069bdfaf-ef1c-f7f2-f9e3-2249bb9dfc41"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-bdb5e23e-0700-20a0-0b21-1ad0901f08b7"
+    }
+  ],
+  "cudaVersion":  "13.0",
+  "writerId":  "bn31pkx91w4z3pgsydsgdnfy5d6scqwz"
+}

epfl_final/SparseAtt/wandb/run-20260708_063331-i61v7k8r/logs/debug-internal.log

@@ -0,0 +1,25 @@
+{"time":"2026-07-08T06:33:31.60933198Z","level":"INFO","msg":"wandb-core"}
+{"time":"2026-07-08T06:33:31.615241845Z","level":"INFO","msg":"stream: starting","core version":"0.26.1"}
+{"time":"2026-07-08T06:33:32.053951715Z","level":"INFO","msg":"stream: created new stream","id":"i61v7k8r"}
+{"time":"2026-07-08T06:33:32.054105415Z","level":"INFO","msg":"handler: started"}
+{"time":"2026-07-08T06:33:32.0565196Z","level":"INFO","msg":"stream: started"}
+{"time":"2026-07-08T06:33:32.05748425Z","level":"INFO","msg":"sender: started"}
+{"time":"2026-07-08T06:33:32.057483018Z","level":"INFO","msg":"writer: started","stream_id":"i61v7k8r"}
+{"time":"2026-07-08T06:33:33.04652623Z","level":"INFO","msg":"filestream: sending request","total_files":1,"console_offset":0,"console_lines":1}
+{"time":"2026-07-08T06:33:33.43908145Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:33:48.046809666Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":0,"events_lines":2,"console_offset":1,"console_lines":1,"uploaded_len":2}
+{"time":"2026-07-08T06:33:48.507722241Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:34:03.04700465Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":2,"events_lines":2}
+{"time":"2026-07-08T06:34:03.387106305Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:34:18.047493623Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":4,"events_lines":2}
+{"time":"2026-07-08T06:34:18.526601778Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:34:33.047269432Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":6,"events_lines":2}
+{"time":"2026-07-08T06:34:33.46690519Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:34:48.047367853Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":8,"events_lines":2}
+{"time":"2026-07-08T06:34:48.40677139Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:35:03.046534231Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":10,"events_lines":2}
+{"time":"2026-07-08T06:35:03.582761178Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:35:18.046991547Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":12,"events_lines":2}
+{"time":"2026-07-08T06:35:18.48837517Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-07-08T06:35:33.047588641Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":14,"events_lines":2}
+{"time":"2026-07-08T06:35:33.495663366Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}