Upload multi-image training dataset

README.md

@@ -1,3 +1,15 @@
 ---
-license: mit
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path: train-*.parquet
 ---
+
+# Multi-image conversation dataset
+
+Each row contains `image` (an ordered list of images) and `conversations`
+(an ordered list of {from, value} objects). Conversation text, whitespace,
+and all image markers are preserved without modification.
+Original image bytes are embedded in Parquet. No resizing or re-encoding
+is performed. Rows and input-image order are preserved.

manifest.json

@@ -0,0 +1,77 @@
+{
+  "format_version": 1,
+  "rows": 66043,
+  "shards": [
+    {
+      "file": "train-00000-of-00010.parquet",
+      "rows": 6620,
+      "size_bytes": 4392236532
+    },
+    {
+      "file": "train-00001-of-00010.parquet",
+      "rows": 6613,
+      "size_bytes": 4391650572
+    },
+    {
+      "file": "train-00002-of-00010.parquet",
+      "rows": 6578,
+      "size_bytes": 4391603216
+    },
+    {
+      "file": "train-00003-of-00010.parquet",
+      "rows": 6598,
+      "size_bytes": 4392000633
+    },
+    {
+      "file": "train-00004-of-00010.parquet",
+      "rows": 6616,
+      "size_bytes": 4392101967
+    },
+    {
+      "file": "train-00005-of-00010.parquet",
+      "rows": 6604,
+      "size_bytes": 4391625802
+    },
+    {
+      "file": "train-00006-of-00010.parquet",
+      "rows": 6598,
+      "size_bytes": 4391014473
+    },
+    {
+      "file": "train-00007-of-00010.parquet",
+      "rows": 6606,
+      "size_bytes": 4391654595
+    },
+    {
+      "file": "train-00008-of-00010.parquet",
+      "rows": 6610,
+      "size_bytes": 4391068299
+    },
+    {
+      "file": "train-00009-of-00010.parquet",
+      "rows": 6600,
+      "size_bytes": 4390325542
+    }
+  ],
+  "features": {
+    "image": {
+      "feature": {
+        "_type": "Image"
+      },
+      "_type": "List"
+    },
+    "conversations": {
+      "feature": {
+        "from": {
+          "dtype": "string",
+          "_type": "Value"
+        },
+        "value": {
+          "dtype": "string",
+          "_type": "Value"
+        }
+      },
+      "_type": "List"
+    }
+  }
+}

train-00000-of-00010.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2be8c7693c200e10d01c3d14d324bd4d13d34aa264e1cab9d99d07cb346bb0c1
+size 4392236532

train-00001-of-00010.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a930ec7c35ae920d945723c2d44581fac81ec0d75ee267355583e8b278bac701
+size 4391650572

train-00002-of-00010.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bbd0278430204099f75f0e16476209934ffbf245a53b8417f177c9e0b7be96a1
+size 4391603216

train-00003-of-00010.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:217752b267aee9d9e534557eee6949da5b829d88ef6715c7dee118297ff17f2e
+size 4392000633

train-00004-of-00010.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:97f0a4861c540e8fb2ea98fb71d1366c483ed3131c6cc36671ef0a1534944e62
+size 4392101967

train-00005-of-00010.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c787b3b7811449430ddca370928ed3ca0861a5b03391b48a6f68966ae7df5df1
+size 4391625802

train-00006-of-00010.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b61bd81b3238d8788d8d0473ff551bd73d7fe6c5bf96fff97fafe8f4f1bd7a4a
+size 4391014473

train-00007-of-00010.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:11e0ae0d1c71b931e5d9d99e41518c302e7a4a68a7396ea4753fd8908c189cd6
+size 4391654595

train-00008-of-00010.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c50ef801c4e7e84dfae80dccf476e4369b7f5a16b820eeedf4a56ac1e5db5138
+size 4391068299

train-00009-of-00010.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a594f010f71eeb1c57fcd21674fe64e905c7a714e8bd5a52c6a7e78825a5be31
+size 4390325542