Add files using upload-large-folder tool

build_report.json

@@ -0,0 +1,302 @@
+{
+  "phase": "channel_align",
+  "seed": 0,
+  "rows": 87656,
+  "rows_by_lang": {
+    "ko": 42050,
+    "en": 45606
+  },
+  "rows_with_voice_prompt": 52036,
+  "rows_with_tool_prompt": 7714,
+  "rows_with_voice_hint": 25660,
+  "rows_prompt_withheld": 9080,
+  "prompt_frames": {
+    "median": 48,
+    "p10": 42,
+    "p90": 48,
+    "below_full": 0.127
+  },
+  "by_lang": {
+    "ko": {
+      "rows": 42050,
+      "rows_with_voice_prompt": 25673,
+      "prompt_share": 0.6105,
+      "rows_with_voice_hint": 12686,
+      "rows_prompt_withheld": 4609,
+      "prompt_frames": {
+        "median": 48,
+        "p10": 31,
+        "p90": 48,
+        "below_full": 0.222
+      }
+    },
+    "en": {
+      "rows": 45606,
+      "rows_with_voice_prompt": 26363,
+      "prompt_share": 0.5781,
+      "rows_with_voice_hint": 12974,
+      "rows_prompt_withheld": 4471,
+      "prompt_frames": {
+        "median": 48,
+        "p10": 48,
+        "p90": 48,
+        "below_full": 0.0345
+      }
+    }
+  },
+  "rows_skipped": {
+    "other_speaker_bleed": 12683
+  },
+  "voice_hint_share_of_prompted": 0.4931,
+  "tool_prompt_share_realized": 0.088,
+  "zone_a_frames_mean": 54.6,
+  "zone_a_share_of_timeline": 0.0761,
+  "opens": {
+    "self": 43682,
+    "user": 43974
+  },
+  "row_frames": {
+    "p10": 156,
+    "median": 796,
+    "p90": 1222,
+    "max": 1498
+  },
+  "rows_in_target_range": 0.5443,
+  "rows_over_hard_cap": 0,
+  "rows_supply_bound": 52225,
+  "frames_by_lang": {
+    "ko": 27739396,
+    "en": 35174319
+  },
+  "frame_share_by_lang": {
+    "en": 0.5591,
+    "ko": 0.4409
+  },
+  "frame_share_declared": {
+    "ko": 0.549,
+    "en": 0.451
+  },
+  "turns": {
+    "p10": 2,
+    "median": 9,
+    "p90": 15,
+    "max": 17
+  },
+  "frames_by_self_supply": {
+    "1": {
+      "p10": 103,
+      "median": 199,
+      "p90": 388,
+      "max": 921,
+      "rows": 26540
+    },
+    "2-3": {
+      "p10": 245,
+      "median": 412,
+      "p90": 662,
+      "max": 1161,
+      "rows": 7174
+    },
+    "4-7": {
+      "p10": 717,
+      "median": 976,
+      "p90": 1254,
+      "max": 1498,
+      "rows": 39100
+    },
+    "8+": {
+      "p10": 774,
+      "median": 1056,
+      "p90": 1327,
+      "max": 1496,
+      "rows": 14842
+    }
+  },
+  "frames_total": 62913715,
+  "hours_of_timeline": 1398.08,
+  "sources": {
+    "ko": [
+      {
+        "name": "kspon",
+        "rows": 587185,
+        "build_report": {
+          "source": "kspon",
+          "reader": "project_amnesty.datasets.kspon.KsponDataset",
+          "repo_commit": "30870612497d014e4f32f4e266fb90df1501bb4c",
+          "cache_version": 2,
+          "rows": 594781,
+          "rows_by_subset": {
+            "train": 587185,
+            "valid": 2398,
+            "test": 5198
+          },
+          "stats": {
+            "total": 627820,
+            "cached": 589573,
+            "encoded": 5208,
+            "failed": 33039
+          },
+          "seconds": 4388.7,
+          "rows_per_second": 135.53,
+          "processor": "/home/work/voice_research/haan/trainer_output/init_scale_match",
+          "align_model": "kresnik/wav2vec2-large-xlsr-korean",
+          "mimi": "kyutai/mimi",
+          "target_lufs": -20.0,
+          "encode_batch_size": 32,
+          "device": "cuda",
+          "source_root": "data/corpora",
+          "limit": null,
+          "python": "3.12.3",
+          "torch": "2.11.0+cu128"
+        }
+      }
+    ],
+    "en": [
+      {
+        "name": "libritts_r",
+        "rows": 149291,
+        "build_report": {
+          "source": "libritts_r",
+          "reader": "project_amnesty.datasets.libritts_r.LibriTTSRDataset",
+          "repo_commit": "5d6df8edeaea5cd77a9d9413930ce781fd791e44",
+          "cache_version": 2,
+          "rows": 159851,
+          "rows_by_subset": {
+            "train": 149291,
+            "test": 4830,
+            "valid": 5730
+          },
+          "stats": {
+            "total": 160267,
+            "cached": 70272,
+            "encoded": 89579,
+            "failed": 416,
+            "drops": {
+              "no_alignable_text": 416
+            }
+          },
+          "seconds": 3721.3,
+          "rows_per_second": 42.96,
+          "processor": "/home/work/voice_research/haan/trainer_output/init_scale_match",
+          "align_model": "facebook/wav2vec2-base-960h",
+          "mimi": "kyutai/mimi",
+          "target_lufs": -20.0,
+          "encode_batch_size": 8,
+          "device": "cuda",
+          "source_root": "data/corpora",
+          "limit": null,
+          "python": "3.12.3",
+          "torch": "2.11.0+cu128"
+        }
+      },
+      {
+        "name": "msp_podcast",
+        "rows": 157363,
+        "build_report": {
+          "source": "msp_podcast",
+          "reader": "project_amnesty.datasets.msp_podcast.MspPodcastDataset",
+          "repo_commit": "d257d9308d235c351e2c00f71ebf9e10b1e92de2",
+          "cache_version": 2,
+          "rows": 245006,
+          "rows_by_subset": {
+            "test": 55539,
+            "train": 157363,
+            "valid": 32104
+          },
+          "stats": {
+            "total": 245273,
+            "cached": 0,
+            "encoded": 245006,
+            "failed": 267,
+            "drops": {
+              "licence_licence": 19014,
+              "licence_no_licence": 3470,
+              "no_transcript": 7,
+              "empty_transcript": 141,
+              "align": 267
+            }
+          },
+          "seconds": 7643.2,
+          "rows_per_second": 32.06,
+          "processor": "/home/work/voice_research/haan/trainer_output/init_scale_match",
+          "align_model": "facebook/wav2vec2-base-960h",
+          "mimi": "kyutai/mimi",
+          "target_lufs": -20.0,
+          "encode_batch_size": 8,
+          "device": "cuda",
+          "source_root": "data/corpora",
+          "limit": null,
+          "python": "3.12.3",
+          "torch": "2.11.0+cu128"
+        }
+      }
+    ]
+  },
+  "settings": {
+    "delay_offsets": [
+      0,
+      1,
+      1,
+      1,
+      1,
+      1,
+      1,
+      1
+    ],
+    "gap_frames": [
+      4,
+      13
+    ],
+    "gap_frames_note": "declared, not measured: turn gaps are Nunchi's subject and modelling them here would write that phase's answer into an earlier phase's data",
+    "skip_other_speaker": true,
+    "system_prompt_pool": {
+      "en": {
+        "plain": 39,
+        "with_tools": 36
+      },
+      "ja": {
+        "plain": 6,
+        "with_tools": 4
+      },
+      "ko": {
+        "plain": 39,
+        "with_tools": 31
+      }
+    },
+    "tool_prompt_share_declared": 0.125,
+    "num_codebooks": 8,
+    "eligibility": {
+      "ko": [
+        {
+          "name": "kspon",
+          "rows": 587185,
+          "rows_label_covered": 442086,
+          "rows_pairable_in_split": 377915
+        }
+      ],
+      "en": [
+        {
+          "name": "libritts_r",
+          "rows": 149291,
+          "rows_label_covered": null,
+          "rows_pairable_in_split": null
+        },
+        {
+          "name": "msp_podcast",
+          "rows": 157363,
+          "rows_label_covered": null,
+          "rows_pairable_in_split": null
+        }
+      ]
+    },
+    "subsets": [
+      "train",
+      "valid",
+      "test"
+    ]
+  },
+  "repo_commit": "b4e5a02f5a9f4f0794dc86fbd139c25abf3b3940",
+  "seconds": 2585.6,
+  "python": "3.12.3",
+  "pipeline_validation": true
+}

shards.json

@@ -0,0 +1,23 @@
+{
+  "arm": "public",
+  "delay_offsets": [
+    0,
+    1,
+    1,
+    1,
+    1,
+    1,
+    1,
+    1
+  ],
+  "rows": {
+    "test": 9660,
+    "train": 24486,
+    "valid": 11460
+  },
+  "shards": {
+    "test": 1,
+    "train": 2,
+    "valid": 1
+  }
+}

test-0000.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8072bd228f803c8f106d9357dcdaad86f6d1501e8260b7d65f0798935aeecded
+size 224563200

train-0000.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:32afda7558dc9991e653680fbd1ce3dbe6de74383889f98699846d2c449399d0
+size 615168000

train-0001.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8d28fa174f3fdacf5da7c39c0534cdff3f08c2e9596c14351ce01db04aca9b64
+size 74158080

valid-0000.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7bbccda397c61a48159a19e9c49070b4fc747f849b64f54d094e07da28459fb8
+size 262000640