issues
search
nerdslab-club
/
cl_model
All models required for curious learner
GNU Affero General Public License v3.0
1
stars
0
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Add cross validation and testing
#45
afmjoaa
opened
4 months ago
0
Update response parser so that result can be easily understood
#44
afmjoaa
opened
6 months ago
0
Fix bleu score calculation
#43
afmjoaa
opened
6 months ago
0
Train and make a perfect model
#42
afmjoaa
opened
6 months ago
0
Integrate Tensorboard to the project
#41
afmjoaa
opened
6 months ago
0
Write a tokenizer to compare the results
#40
afmjoaa
opened
6 months ago
1
#38: Gpu trainable
#39
Prattoy
closed
7 months ago
0
Complete the training and get the result.
#38
afmjoaa
closed
6 months ago
1
BUG: CL-Pre-Trainer
#36
afmjoaa
closed
7 months ago
1
Write data loader for GPU training.
#37
afmjoaa
opened
7 months ago
4
Write proper documentation
#35
afmjoaa
opened
8 months ago
0
Train and evaluate task specific models performance on the specific tasks on GPU
#34
afmjoaa
opened
9 months ago
0
Make another model(Transformer) for task specific training(POC)
#33
afmjoaa
opened
9 months ago
0
Train the pre-trainer model using GPU on the scoped data
#32
afmjoaa
closed
7 months ago
2
Create tests for evaluating model performence **
#31
afmjoaa
closed
7 months ago
2
Complete training testing and validation data scope
#30
afmjoaa
closed
8 months ago
1
Explore options for better training results with less time
#29
afmjoaa
closed
8 months ago
3
#26 fix training and inference
#28
afmjoaa
closed
8 months ago
0
Explore all possible solution to improve the training and loading of the saved model.
#27
afmjoaa
closed
10 months ago
1
Implement proper inference method using the saved model.
#26
afmjoaa
closed
8 months ago
5
Implement the cl-pretrainer training and inference class
#25
afmjoaa
closed
10 months ago
0
Implement the cl-pretrainer class
#24
afmjoaa
closed
11 months ago
0
#11: Implement the category router to figure out the classification head for output token
#23
afmjoaa
closed
11 months ago
0
Modify the #9 ticket output to the current proposed architecture for task based work.
#22
afmjoaa
opened
11 months ago
0
#10: Complete pre-trainer model category-map probability prediction.
#21
afmjoaa
closed
11 months ago
0
#9 add embedding manager
#20
afmjoaa
closed
11 months ago
0
#10 category map probability
#19
afmjoaa
closed
11 months ago
0
#8: Corpus source name and tgt name changed for readability
#18
Prattoy
closed
11 months ago
0
#8: Use ALiBiBi instead of PE in the pre-trainer model
#17
afmjoaa
closed
11 months ago
0
#5 and #6 save and load checkpoint
#16
afmjoaa
closed
11 months ago
0
Write the response parser.
#15
afmjoaa
closed
8 months ago
1
Complete pre-trainer model output probability prediction with back propagation.
#14
afmjoaa
opened
11 months ago
0
Implement the cl-pretrainer training
#13
afmjoaa
closed
10 months ago
2
Implement the cl-pretrainer class
#12
afmjoaa
closed
11 months ago
1
Implement the category router to figure out the classification head for output token
#11
afmjoaa
closed
11 months ago
1
Complete pre-trainer model category-map probability prediction with back propagation.
#10
afmjoaa
closed
11 months ago
2
Use combined embeddings instead of random nn.embeddings, tokenizer and vocubulary builder.
#9
afmjoaa
closed
11 months ago
1
Use Alibibi instead of PE in the pre-trainer model.
#8
afmjoaa
closed
11 months ago
2
Draw a detailed architectural picture for the pre-trainer model
#7
afmjoaa
closed
11 months ago
1
Resume from a checkpoint using modular saved weights.
#6
afmjoaa
opened
12 months ago
1
Save current transformer model segment wise
#5
afmjoaa
opened
12 months ago
1
Complete the transformer from scratch documentation
#4
afmjoaa
closed
12 months ago
1
Complete the embeddings manager.
#3
afmjoaa
closed
11 months ago
3
complete the categorical and task encoder and ALiBiBi Encoder
#2
afmjoaa
closed
1 year ago
1
Complete positional embedding class of embeddings manager package
#1
afmjoaa
closed
1 year ago
1