Manekineko's workspace
Runs
51
Name
1 visualized
State
Notes
User
Tags
Created
Runtime
Sweep
adam_betas
adam_offload
apply_chat_template
aux_loss_coef
bf16
ckpt_path
dataset
dataset_probs
disable_ds_ckpt
disable_fast_tokenizer
eval_split
eval_steps
flash_attn
full_determinism
gradient_checkpointing
gradient_checkpointing_use_reentrant
input_key
input_template
kd_coef
l2
learning_rate
load_checkpoint
load_in_4bit
local_rank
logging_steps
lora_alpha
lora_dropout
lora_rank
lr_scheduler
lr_warmup_ratio
max_ckpt_mem
max_ckpt_num
max_epochs
max_len
max_norm
max_samples
micro_train_batch_size
multiturn
output_key
overlap_comm
packing_samples
pretrain
pretrain_mode
ring_attn_size
Crashed
-
randomfoo
18h 11m 1s
-
[0.9,0.95]
false
true
0
true
/fsx2/checkpoints
/fsx/ubuntu/meti/data/openrlhf/sft.geniac.jsonl
1.0
false
false
test
-1
true
false
true
false
conversations
User: {}
Assistant:
-
0
0.00000264
true
false
-1
1
16
0
0
cosine_with_min_lr
0.05
100000000
1
3
4096
1
1000000
1
false
-
true
true
meta-llama/Llama-3.1-405B-Instruct
false
8
Crashed
-
randomfoo
9d 20h 35m 26s
-
[0.9,0.95]
false
true
0
true
/fsx2/checkpoints
/fsx/ubuntu/meti/data/openrlhf/sft.geniac.jsonl
1.0
false
false
test
-1
true
false
true
false
conversations
User: {}
Assistant:
-
0
0.00000264
false
false
-1
1
16
0
0
cosine_with_min_lr
0.05
100000000
1
3
4096
1
1000000
1
false
-
true
true
meta-llama/Llama-3.1-405B-Instruct
false
8
Crashed
-
randomfoo
54m
-
[0.9,0.95]
false
true
0
true
/fsx2/checkpoints
/fsx/ubuntu/meti/data/openrlhf/sft.geniac.jsonl
1.0
false
false
test
-1
true
false
true
false
conversations
User: {}
Assistant:
-
0
0.00000264
false
false
-1
1
16
0
0
cosine_with_min_lr
0.05
100000000
1
3
4096
1
1000000
1
false
-
true
true
meta-llama/Llama-3.1-405B-Instruct
false
8
Crashed
-
randomfoo
3d 9h 46m 31s
-
[0.9,0.95]
false
true
0
true
/fsx2/checkpoints
/fsx/ubuntu/meti/data/openrlhf/sft.geniac.jsonl
1.0
false
false
test
-1
true
false
true
false
conversations
User: {}
Assistant:
-
0
0.0000012
false
false
-1
1
16
0
0
cosine_with_min_lr
0.05
100000000
1
3
4096
1
1000000
1
false
-
true
true
/fsx/ubuntu/.cache/huggingface/hub/models--ubitus--alice-v3-1-405b-instruct-rc1/snapshots/34c1d561cec92bf3e96c86244f0ef15f740cdc1a
false
8
Finished
-
randomfoo
7h 6m 47s
-
[0.9,0.95]
false
true
0
true
./ckpt/checkpoints_sft
shisa-ai/paradox_test_set_200k_sharegpt-easyOpenRLHFformat
1.0
false
false
test
-1
true
false
true
false
question
User: {}
Assistant:
-
0
0.000001
false
false
-1
1
16
0
0
cosine_with_min_lr
0.05
100000000
3
3
8192
1
100000
4
false
response
true
true
google/gemma-3-27b-it
false
2
Crashed
-
randomfoo
11m 46s
-
[0.9,0.95]
false
true
0
true
./ckpt/checkpoints_sft
shisa-ai/paradox_test_set_200k_sharegpt-easyOpenRLHFformat
1.0
false
false
test
-1
true
false
true
false
question
User: {}
Assistant:
-
0
0.000002
false
false
-1
1
16
0
0
cosine_with_min_lr
0.05
100000000
3
3
8192
1
100000
4
false
response
true
true
google/gemma-3-27b-it
false
2
Finished
-
randomfoo
2h 57m 47s
-
[0.9,0.95]
false
true
0
true
./ckpt/checkpoints_sft
shisa-ai/paradox_test_set_200k_sharegpt-easyOpenRLHFformat
1.0
false
false
test
-1
true
false
true
false
question
User: {}
Assistant:
-
0
0.000002
false
false
-1
1
16
0
0
cosine_with_min_lr
0.05
100000000
3
3
8192
1
50000
4
false
response
true
true
google/gemma-3-12b-it
false
2
Finished
-
randomfoo
1h 13m 36s
-
[0.9,0.95]
false
true
0
true
./ckpt/checkpoints_sft
shisa-ai/paradox_test_set_200k_sharegpt-easyOpenRLHFformat
1.0
false
false
test
-1
true
false
true
false
question
User: {}
Assistant:
-
0
0.0000048
false
false
-1
1
16
0
0
cosine_with_min_lr
0.05
100000000
3
3
8192
1
50000
4
false
response
true
true
google/gemma-3-4b-it
false
2
Finished
-
randomfoo
44m 18s
-
[0.9,0.95]
false
true
0
true
./ckpt/checkpoints_sft
shisa-ai/paradox_test_set_200k_sharegpt-easyOpenRLHFformat
1.0
false
false
test
-1
true
false
true
false
question
User: {}
Assistant:
-
0
0.0000068
false
false
-1
1
16
0
0
cosine_with_min_lr
0.05
100000000
3
3
8192
1
30000
4
false
response
true
true
google/gemma-3-4b-it
false
2
Finished
-
randomfoo
16h 15m 41s
-
[0.9,0.95]
false
true
0
true
./ckpt/checkpoints_sft
/fsx/ubuntu/meti/data/openrlhf/sft.shisa-v2.jsonl
1.0
false
false
test
-1
true
false
true
false
conversations
User: {}
Assistant:
-
0
0.0000068
false
false
-1
1
16
0
0
cosine_with_min_lr
0.05
100000000
3
3
8192
1
1000000
4
false
-
true
true
meta-llama/Llama-3.3-70B-Instruct
false
2
Finished
-
randomfoo
1h 40m 51s
-
[0.9,0.95]
false
false
0
true
./ckpt/checkpoints_kd
shisa-ai/translation_distil_set_4096_tokens
1.0
-
false
test
-1
true
false
true
false
question
User: {}
Assistant:
0.4
0
0.000005
false
false
-1
1
16
0
0
cosine_with_min_lr
0.05
100000000
3
3
4096
1
1000000
2
-
response
true
-
/fsx2/outputs/ablation-151-finalsft-shisa-v2-mistral-nemo-12b
false
-
Crashed
-
randomfoo
16s
-
[0.9,0.95]
false
false
0
true
./ckpt/checkpoints_kd
shisa-ai/translation_distil_set_4096_tokens
1.0
-
false
test
-1
true
false
true
false
question
User: {}
Assistant:
0.4
0
0.000005
false
false
-1
1
16
0
0
cosine_with_min_lr
0.05
100000000
3
3
4096
1
1000000
4
-
response
true
-
/fsx2/outputs/ablation-151-finalsft-shisa-v2-mistral-nemo-12b
false
-
Finished
-
randomfoo
31m 46s
-
[0.9,0.95]
false
false
0
true
./ckpt/checkpoints_kd
shisa-ai/translation_distil_set_2048_tokens
1.0
-
false
test
-1
true
false
true
false
question
User: {}
Assistant:
0.4
0
0.000005
false
false
-1
1
16
0
0
cosine_with_min_lr
0.05
100000000
3
3
4096
1
1000000
4
-
response
true
-
/fsx2/outputs/ablation-151-finalsft-shisa-v2-mistral-nemo-12b
false
-
Finished
8n:70b ds=0.15.0 GC rsz=2 rstr=2
randomfoo
15h 39m 8s
-
[0.9,0.95]
false
true
0
true
./ckpt/checkpoints_sft
/fsx/ubuntu/meti/data/openrlhf/sft.shisa-v2.jsonl
1.0
false
false
test
-1
true
false
true
false
conversations
User: {}
Assistant:
-
0
0.0000068
false
false
-1
1
16
0
0
cosine_with_min_lr
0.05
100000000
3
3
8192
1
1000000
4
false
-
true
true
meta-llama/Llama-3.3-70B-Instruct
false
2
Crashed
-
randomfoo
9m 16s
-
[0.9,0.95]
false
true
0
true
./ckpt/checkpoints_sft
/fsx/ubuntu/meti/data/openrlhf/sft.shisa-v2.jsonl
1.0
false
false
test
-1
true
false
true
false
conversations
User: {}
Assistant:
-
0
0.0000068
false
false
-1
1
16
0
0
cosine_with_min_lr
0.05
100000000
3
3
8192
1
1000000
4
false
-
true
true
meta-llama/Llama-3.3-70B-Instruct
false
4
Crashed
-
randomfoo
11m 31s
-
[0.9,0.95]
false
true
0
true
./ckpt/checkpoints_sft
/fsx/ubuntu/meti/data/openrlhf/sft.shisa-v2.jsonl
1.0
false
false
test
-1
true
false
true
false
conversations
User: {}
Assistant:
-
0
0.0000068
false
false
-1
1
16
0
0
cosine_with_min_lr
0.05
100000000
3
3
8192
1
1000000
4
false
-
true
true
meta-llama/Llama-3.3-70B-Instruct
false
8
Crashed
-
randomfoo
45m 1s
-
[0.9,0.95]
false
true
0
true
./ckpt/checkpoints_sft
/fsx/ubuntu/meti/data/openrlhf/sft.shisa-v2.jsonl
1.0
false
false
test
-1
true
false
false
false
conversations
User: {}
Assistant:
-
0
0.0000068
false
false
-1
1
16
0
0
cosine_with_min_lr
0.05
100000000
3
3
8192
1
1000000
4
false
-
true
true
meta-llama/Llama-3.3-70B-Instruct
false
8
Crashed
-
randomfoo
4h 45m 1s
-
[0.9,0.95]
false
true
0
true
./ckpt/checkpoints_sft
/fsx/ubuntu/meti/data/openrlhf/sft.shisa-v2.jsonl
1.0
false
false
test
-1
true
false
false
false
conversations
User: {}
Assistant:
-
0
0.0000068
false
false
-1
1
16
0
0
cosine_with_min_lr
0.05
100000000
3
3
6000
1
1000000
2
false
-
true
true
meta-llama/Llama-3.3-70B-Instruct
false
4
Crashed
-
randomfoo
26m 45s
-
[0.9,0.95]
false
true
0
true
./ckpt/checkpoints_sft
/fsx/ubuntu/meti/data/openrlhf/sft.shisa-v2.jsonl
1.0
false
false
test
-1
true
false
false
false
conversations
User: {}
Assistant:
-
0
0.0000068
false
false
-1
1
16
0
0
cosine_with_min_lr
0.05
100000000
3
3
8192
1
1000000
2
false
-
true
true
meta-llama/Llama-3.3-70B-Instruct
false
4
Crashed
-
randomfoo
3m 17s
-
[0.9,0.95]
false
true
0
true
./ckpt/checkpoints_sft
/fsx/ubuntu/meti/data/openrlhf/sft.shisa-v2.jsonl
1.0
false
false
test
-1
true
false
false
false
conversations
User: {}
Assistant:
-
0
0.000006
false
false
-1
1
16
0
0
cosine_with_min_lr
0.05
100000000
3
3
8192
1
1000000
2
false
-
true
true
meta-llama/Llama-3.3-70B-Instruct
false
8
1-20
of 51