issues
search
ReaLLMASIC
/
nanoGPT
The simplest, fastest repository for training/finetuning medium-sized GPTs.
MIT License
23
stars
17
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
added lower_triangle
#261
mmoffatt2
opened
1 day ago
0
Lower triangle
#260
mmoffatt2
closed
1 day ago
0
Make softmax io logging more uniform
#259
gkielian
closed
6 days ago
0
Add factorization support, steering support, and eval only support with longer contexts
#258
gkielian
closed
6 days ago
1
Add code for finetuning with FIRE
#257
gkielian
opened
1 week ago
0
modify overflow condition to greater than 88
#256
Hrancheng
closed
1 week ago
0
Update sample.py with eval parameters
#255
gkielian
closed
1 week ago
1
Add update block size to train.py for evaluation
#254
gkielian
closed
1 week ago
1
add softmax variant IO interval
#253
Hrancheng
closed
6 days ago
0
Add MLP Expansion factor control and sweep
#252
gkielian
closed
1 week ago
0
Fixed One Bug in FIRE - PR #246 v4
#251
Mars-Cat2023
closed
1 week ago
0
Merge Factorization
#250
gkielian
closed
6 days ago
2
Add numpy mapping
#249
klei22
opened
2 weeks ago
0
add sliding windows using flex attention and attn_gym packages
#248
Hrancheng
opened
2 weeks ago
1
Add softrelumax fire sweep
#247
gkielian
opened
2 weeks ago
0
Parameterized FIRE (Adding Options for FIRE)
#246
Mars-Cat2023
closed
2 weeks ago
2
Add softmax sweep to benchmark softmaxes vs context
#245
klei22
closed
2 weeks ago
1
Organize starting directory
#244
klei22
closed
2 weeks ago
0
Add progress bar to train.py
#243
gkielian
closed
2 weeks ago
0
upload current progress about tokens for snac_text
#242
xinyixuu
opened
2 weeks ago
0
Add option to get sample inference after each val
#241
gkielian
closed
2 weeks ago
1
Gptconfig fix
#240
djlisbonne
closed
2 weeks ago
0
Add progress bar and eta to train.py
#239
gkielian
closed
2 weeks ago
2
Huggingface Model Sample and Upload
#238
mmoffatt2
closed
3 weeks ago
0
Huggingface Model
#237
mmoffatt2
closed
3 weeks ago
0
Add numpy hw test
#236
gkielian
closed
2 weeks ago
0
Add ConSmaxV2 -- per head gamma beta and per gamma/beta scaling factor
#235
gkielian
closed
2 weeks ago
2
Add conditions for softmax io logging
#234
gkielian
closed
2 weeks ago
1
Hf from_pretrained fix
#233
djlisbonne
closed
3 weeks ago
0
Muti parameter/objective in run_vizier.py
#232
Hrancheng
opened
3 weeks ago
3
Add ReLUMax variation and sweep
#231
gkielian
closed
3 weeks ago
0
Moved all train.py statistics functions to new folder
#230
mmoffatt2
closed
4 weeks ago
0
Added Quantization Granularity of Matmul Inputs
#229
mmoffatt2
closed
3 weeks ago
0
Experiments: Rotary Position Embedding Swapping/Finetuning for HW Effiiciency
#228
gkielian
opened
1 month ago
0
modified whisper_snac.sh to run the whole process
#227
xinyixuu
closed
3 weeks ago
0
MoE expert sharing and freezing support
#226
djlisbonne
opened
1 month ago
0
Add implementation of Rotary Embeddings
#225
klei22
closed
1 month ago
3
Add ability to save quantized weights/activations, scale factors, and zero points
#224
mmoffatt2
closed
3 weeks ago
1
Add model param section
#223
gkielian
closed
1 month ago
0
Quest: Add Benchmarks and Pretokenized JSON for training
#222
gkielian
opened
1 month ago
0
Exploration Augmentation: Add means to pause and resume search
#221
gkielian
opened
1 month ago
0
Exploration Augmentation: Run_experiments Best Results Log
#220
gkielian
opened
1 month ago
0
Exploration Automation: Obtain Dataset, Select Tokenization, and Cache results
#219
gkielian
opened
1 month ago
0
Whisper script path adjustments
#218
gkielian
closed
1 month ago
0
Revert 1 add snac tokens patch
#217
xinyixuu
closed
1 month ago
0
Quantized Input/Output Activations
#216
mmoffatt2
closed
1 month ago
0
Replace counters with flag for monitoring recomputed
#215
gkielian
closed
1 month ago
0
Quantized Linear
#214
mmoffatt2
closed
1 month ago
2
Sample.py: fix inference for p50k_base and cl100k_base
#213
gkielian
opened
1 month ago
0
Added partial code for snac tokens
#212
xinyixuu
closed
1 month ago
7
Next