Outer Step 0. Inner Step 0. Batch Size 0
Browse files- config.json +4 -183
- model.safetensors +2 -2
config.json
CHANGED
@@ -1,5 +1,5 @@
|
|
1 |
{
|
2 |
-
"_name_or_path": "distributed/optimized-gpt2-1b-
|
3 |
"activation_function": "gelu_new",
|
4 |
"all_reduce_scores": {
|
5 |
"0": "NON_PARTICIPATING",
|
@@ -264,189 +264,10 @@
|
|
264 |
],
|
265 |
"attn_pdrop": 0.1,
|
266 |
"auto_map": {
|
267 |
-
"AutoConfig": "distributed/optimized-gpt2-
|
268 |
-
"AutoModelForCausalLM": "distributed/optimized-gpt2-
|
269 |
},
|
270 |
-
"block_list": [
|
271 |
-
3846514,
|
272 |
-
3846514,
|
273 |
-
3846514,
|
274 |
-
3846514,
|
275 |
-
3846514,
|
276 |
-
3846514,
|
277 |
-
3846514,
|
278 |
-
3846514,
|
279 |
-
3846514,
|
280 |
-
3846514,
|
281 |
-
3846514,
|
282 |
-
3846514,
|
283 |
-
3846514,
|
284 |
-
3846514,
|
285 |
-
3846514,
|
286 |
-
3846514,
|
287 |
-
3846514,
|
288 |
-
3846514,
|
289 |
-
3846514,
|
290 |
-
3846514,
|
291 |
-
3846514,
|
292 |
-
3846514,
|
293 |
-
3846514,
|
294 |
-
3846514,
|
295 |
-
3846514,
|
296 |
-
3846514,
|
297 |
-
3846514,
|
298 |
-
3846514,
|
299 |
-
3846514,
|
300 |
-
3846514,
|
301 |
-
3846514,
|
302 |
-
3846514,
|
303 |
-
3846514,
|
304 |
-
3846514,
|
305 |
-
3846514,
|
306 |
-
3846514,
|
307 |
-
3846514,
|
308 |
-
3846514,
|
309 |
-
3846514,
|
310 |
-
3846514,
|
311 |
-
3846514,
|
312 |
-
3846514,
|
313 |
-
3846514,
|
314 |
-
3846514,
|
315 |
-
3846514,
|
316 |
-
3846514,
|
317 |
-
3846514,
|
318 |
-
3846514,
|
319 |
-
3846514,
|
320 |
-
3846514,
|
321 |
-
3846514,
|
322 |
-
3846514,
|
323 |
-
3846514,
|
324 |
-
3846514,
|
325 |
-
3846514,
|
326 |
-
3846514,
|
327 |
-
3846514,
|
328 |
-
3846514,
|
329 |
-
3846514,
|
330 |
-
3846514,
|
331 |
-
3846514,
|
332 |
-
3846514,
|
333 |
-
3846514,
|
334 |
-
3846514,
|
335 |
-
3846514,
|
336 |
-
3846514,
|
337 |
-
3846514,
|
338 |
-
3846514,
|
339 |
-
3846514,
|
340 |
-
3846514,
|
341 |
-
3846514,
|
342 |
-
3846514,
|
343 |
-
3846514,
|
344 |
-
3846514,
|
345 |
-
3846514,
|
346 |
-
3846514,
|
347 |
-
3846514,
|
348 |
-
3846514,
|
349 |
-
3846514,
|
350 |
-
3846514,
|
351 |
-
3846514,
|
352 |
-
3846514,
|
353 |
-
3846514,
|
354 |
-
3846514,
|
355 |
-
3846514,
|
356 |
-
3846514,
|
357 |
-
3846514,
|
358 |
-
3846514,
|
359 |
-
3846514,
|
360 |
-
3846514,
|
361 |
-
3846514,
|
362 |
-
3846514,
|
363 |
-
3846514,
|
364 |
-
3846514,
|
365 |
-
3846514,
|
366 |
-
3846514,
|
367 |
-
3846514,
|
368 |
-
3846514,
|
369 |
-
3846514,
|
370 |
-
3846514,
|
371 |
-
3846514,
|
372 |
-
3846514,
|
373 |
-
3846514,
|
374 |
-
3846514,
|
375 |
-
3846514,
|
376 |
-
3846514,
|
377 |
-
3846514,
|
378 |
-
3846514,
|
379 |
-
3846514,
|
380 |
-
3846514,
|
381 |
-
3846514,
|
382 |
-
3846514,
|
383 |
-
3846514,
|
384 |
-
3846514,
|
385 |
-
3846514,
|
386 |
-
3846514,
|
387 |
-
3846514,
|
388 |
-
3846514,
|
389 |
-
3846514,
|
390 |
-
3846514,
|
391 |
-
3846514,
|
392 |
-
3846514,
|
393 |
-
3846514,
|
394 |
-
3846514,
|
395 |
-
3846514,
|
396 |
-
3846514,
|
397 |
-
3846514,
|
398 |
-
3846514,
|
399 |
-
3846514,
|
400 |
-
3846514,
|
401 |
-
3846514,
|
402 |
-
3846514,
|
403 |
-
3846514,
|
404 |
-
3846514,
|
405 |
-
3846514,
|
406 |
-
3846514,
|
407 |
-
3846514,
|
408 |
-
3846514,
|
409 |
-
3846514,
|
410 |
-
3846514,
|
411 |
-
3846514,
|
412 |
-
3846514,
|
413 |
-
3846514,
|
414 |
-
3846514,
|
415 |
-
3846514,
|
416 |
-
3846514,
|
417 |
-
3846514,
|
418 |
-
3846514,
|
419 |
-
3846514,
|
420 |
-
3846514,
|
421 |
-
3846514,
|
422 |
-
3846514,
|
423 |
-
3846514,
|
424 |
-
3846514,
|
425 |
-
3846514,
|
426 |
-
3846514,
|
427 |
-
3846514,
|
428 |
-
3846514,
|
429 |
-
3846514,
|
430 |
-
3846514,
|
431 |
-
3846514,
|
432 |
-
3846514,
|
433 |
-
3846514,
|
434 |
-
3846514,
|
435 |
-
3846514,
|
436 |
-
3846514,
|
437 |
-
3846514,
|
438 |
-
3846514,
|
439 |
-
3846514,
|
440 |
-
3846514,
|
441 |
-
3846514,
|
442 |
-
3846514,
|
443 |
-
3846514,
|
444 |
-
3846514,
|
445 |
-
3846514,
|
446 |
-
3846514,
|
447 |
-
3846514,
|
448 |
-
3846514
|
449 |
-
],
|
450 |
"block_size": 1024,
|
451 |
"bos_token_id": 50256,
|
452 |
"embd_pdrop": 0.1,
|
|
|
1 |
{
|
2 |
+
"_name_or_path": "distributed/optimized-gpt2-1b-vtestnet-v1",
|
3 |
"activation_function": "gelu_new",
|
4 |
"all_reduce_scores": {
|
5 |
"0": "NON_PARTICIPATING",
|
|
|
264 |
],
|
265 |
"attn_pdrop": 0.1,
|
266 |
"auto_map": {
|
267 |
+
"AutoConfig": "distributed/optimized-gpt2-500m--configuration_gpt_optimized.GPTOptimConfig",
|
268 |
+
"AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
|
269 |
},
|
270 |
+
"block_list": [],
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
271 |
"block_size": 1024,
|
272 |
"bos_token_id": 50256,
|
273 |
"embd_pdrop": 0.1,
|
model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c8a3d712b2c6e4bc5f540e5a3f1a009c9e4b1f8606024d9ec8499429e6e14214
|
3 |
+
size 4040701744
|