Datasets:

einarolafsson
/

spacr-tutorials

schema int64	language string	lang_code string	dialect string	voice string	speed float64	sample_rate int64	total_duration float64	media_sha256 string	media_bytes int64	render_fingerprint string	render_inputs dict	scenes list
1	en	a	us	af_aoede	0.96	24,000	101.495	085c87ab1abd7f56d50b1892458429e69688c86cf4220df96f886f15b2a43605	650,636	0214594905063b08c5acb5be9ec731057398c0471a88baa55d52e19984ac8330	{ "renderer": "render_all_voices-v4", "lesson": "19_train_cellpose", "language": "en", "lang_code": "a", "dialect": "us", "voice": "af_aoede", "speed": 0.96, "scenes": [ { "narration": "Train Cellpose fits a segmentation model from paired microscopy images and instance-label masks. The console...	[ { "scene": 1, "speech_start": 0, "speech_end": 12.445, "scene_end": 12.965, "duration": 12.965, "text": "Train Cellpose fits a segmentation model from paired microscopy images and instance-label masks. The console reports data validation, progress, checkpoint path, and reproducibility record...
1	en	a	us	af_bella	1.03	24,000	106.625	71147fa37e27c4d379fd56534b1857faae78b2a181de4b1acebf4bfb7ad643f9	701,301	648b152f65d95fbf8a2c24f2363cbad33505997fb356d3891d670578decd4feb	{ "renderer": "render_all_voices-v4", "lesson": "19_train_cellpose", "language": "en", "lang_code": "a", "dialect": "us", "voice": "af_bella", "speed": 1.03, "scenes": [ { "narration": "Train Cellpose fits a segmentation model from paired microscopy images and instance-label masks. The console...	[ { "scene": 1, "speech_start": 0, "speech_end": 13.17, "scene_end": 13.69, "duration": 13.69, "text": "Train Cellpose fits a segmentation model from paired microscopy images and instance-label masks. The console reports data validation, progress, checkpoint path, and reproducibility records."...
1	en	a	us	af_heart	1	24,000	104.235	c4b245b68a390029f58a1b44ad25ab5cec508422fbc8a7014ef2806aabb6eb10	669,704	7f2acf10467788f23c22bb80073c14f3cd0716384374cec9414b63404dc31b22	{ "renderer": "render_all_voices-v4", "lesson": "19_train_cellpose", "language": "en", "lang_code": "a", "dialect": "us", "voice": "af_heart", "speed": 1, "scenes": [ { "narration": "Train Cellpose fits a segmentation model from paired microscopy images and instance-label masks. The console re...	[ { "scene": 1, "speech_start": 0, "speech_end": 12.74, "scene_end": 13.26, "duration": 13.26, "text": "Train Cellpose fits a segmentation model from paired microscopy images and instance-label masks. The console reports data validation, progress, checkpoint path, and reproducibility records."...
1	en	a	us	af_jessica	0.88	24,000	98.985	fdef8462d0a73efa25a49a443e87f3911c098f45e805a60a2df5638b036f0ce9	637,679	415a570a06990d72ef7832ba780c42a2e741f5766029ce1485089e061bb4d21f	{ "renderer": "render_all_voices-v4", "lesson": "19_train_cellpose", "language": "en", "lang_code": "a", "dialect": "us", "voice": "af_jessica", "speed": 0.88, "scenes": [ { "narration": "Train Cellpose fits a segmentation model from paired microscopy images and instance-label masks. The conso...	[ { "scene": 1, "speech_start": 0, "speech_end": 11.95, "scene_end": 12.469999999999999, "duration": 12.469999999999999, "text": "Train Cellpose fits a segmentation model from paired microscopy images and instance-label masks. The console reports data validation, progress, checkpoint path, and...
1	en	a	us	af_river	0.88	24,000	99.895	2c99874c2f3873db142d00eeb22688505e8d1f51512cc8e4fe9b8dbf7d67505f	640,967	45b242440aa07825c690626ea33164b17877c127e3359962b41c8eb9866deb3f	{ "renderer": "render_all_voices-v4", "lesson": "19_train_cellpose", "language": "en", "lang_code": "a", "dialect": "us", "voice": "af_river", "speed": 0.88, "scenes": [ { "narration": "Train Cellpose fits a segmentation model from paired microscopy images and instance-label masks. The console...	[ { "scene": 1, "speech_start": 0, "speech_end": 11.98, "scene_end": 12.5, "duration": 12.5, "text": "Train Cellpose fits a segmentation model from paired microscopy images and instance-label masks. The console reports data validation, progress, checkpoint path, and reproducibility records.", ...
1	en	a	us	af_sarah	0.97	24,000	104.555	8153cff68d0fd27fffb42df6515a621f837dd61fd13ad05f9f3a58fe8578f738	694,385	24b72f1d2e90a0bea80e214ee2d377302487f1c96a443abe0b3f3d40f46e9eb6	{ "renderer": "render_all_voices-v4", "lesson": "19_train_cellpose", "language": "en", "lang_code": "a", "dialect": "us", "voice": "af_sarah", "speed": 0.97, "scenes": [ { "narration": "Train Cellpose fits a segmentation model from paired microscopy images and instance-label masks. The console...	[ { "scene": 1, "speech_start": 0, "speech_end": 12.765, "scene_end": 13.285, "duration": 13.285, "text": "Train Cellpose fits a segmentation model from paired microscopy images and instance-label masks. The console reports data validation, progress, checkpoint path, and reproducibility record...
1	en	a	us	af_sky	0.96	24,000	105.265	4232fdc3ba50e2678d88c543af86913f0e670b7e61f8963b801cdf411d6db188	674,386	8ab04993dcef90f340f9063aa32710e61fe4946232cbbf28497ddc1e870b3c68	{ "renderer": "render_all_voices-v4", "lesson": "19_train_cellpose", "language": "en", "lang_code": "a", "dialect": "us", "voice": "af_sky", "speed": 0.96, "scenes": [ { "narration": "Train Cellpose fits a segmentation model from paired microscopy images and instance-label masks. The console r...	[ { "scene": 1, "speech_start": 0, "speech_end": 12.66, "scene_end": 13.18, "duration": 13.18, "text": "Train Cellpose fits a segmentation model from paired microscopy images and instance-label masks. The console reports data validation, progress, checkpoint path, and reproducibility records."...
1	en	a	us	am_adam	0.95	24,000	102.965	355cf11e8eb34dc24f39d2b8dbe6159c8ada1d09634d2c383533cf17ad4e3397	673,247	75aa018f3200865379f73dd16c838a11db65bbf64635a01728e679745cfcfba8	{"renderer":"render_all_voices-v4","lesson":"19_train_cellpose","language":"en","lang_code":"a","dia(...TRUNCATED)	[{"scene":1,"speech_start":0.0,"speech_end":12.87,"scene_end":13.389999999999999,"duration":13.38999(...TRUNCATED)
1	en	a	us	am_echo	0.94	24,000	106.715	c416c77d637308c4f949b6df535d8e99085a796b80559ab582a8856a9923bf9c	691,151	382c0fbddfc51c964d206ef2d1efad3e4c46dc85306b16754e8d9c67458a99e0	{"renderer":"render_all_voices-v4","lesson":"19_train_cellpose","language":"en","lang_code":"a","dia(...TRUNCATED)	[{"scene":1,"speech_start":0.0,"speech_end":12.845,"scene_end":13.365,"duration":13.365,"text":"Trai(...TRUNCATED)
1	en	a	us	am_eric	0.88	24,000	98.71	d40a8f948ae3699b248957d7fc3ba829240366df1a73a519bbed3c23b0acd043	620,195	0a64555bcfd10aaa521a0e2f5b7a484227999a39c9e470bf4ec030fba9193230	{"renderer":"render_all_voices-v4","lesson":"19_train_cellpose","language":"en","lang_code":"a","dia(...TRUNCATED)	[{"scene":1,"speech_start":0.0,"speech_end":11.97,"scene_end":12.49,"duration":12.49,"text":"Train C(...TRUNCATED)

End of preview. Expand in Data Studio

spaCR tutorial media

Narration and 4K video for the spaCR interactive tutorial library, served directly to https://einarolafsson.github.io/spacr/tutorials/.

spaCR is a toolkit for microscopy and single-cell analysis of pooled CRISPR screens. This repository holds the media its 40-lesson tutorial player streams; it is not a training dataset.

Why it lives here

GitHub Pages caps a published site at 1 GB. The full narration set is 2,662 MiB across 54 voices, so the site could only ever publish one voice per language — 27 of the 28 English voices were unreachable. Hosting the audio here removes the cap rather than rationing it, and leaves room for the 4K video too.

Layout

Paths mirror the tutorial player's own, so a single base URL resolves both:

<lesson_id>/audio/<language>/<voice>.m4a    narration
<lesson_id>/audio/<language>/<voice>.json   word timings for captions and sync
<lesson_id>/video/<lesson_id>_silent.mp4    4K silent master (3840x2160)

For example:

07_mask/audio/en/af_heart.m4a
07_mask/audio/ja/jf_alpha.m4a
07_mask/video/07_mask_silent.mp4


Lessons	40
Narrated languages	8 (en, es, fr, hi, it, pt-BR, ja, zh-CN)
Voices	54
Narration files	2,160 (~2,662 MiB)
4K silent masters	40 (~681 MiB)

How the player uses it

The video carries no audio track. The player plays a silent master and overlays whichever narration voice the visitor picks, mapping the two timelines onto each other so any voice fits any lesson. The site offers 1440p by default and the 4K cut from here as a quality option; both are the same cut on the same timeline, so switching quality does not interrupt narration.

Narration was rendered with Kokoro-82M.