Tried depth-upscaling Cosmo-1b by duplicating 6 layers, then LISA-training on a dataset reasonably similar to the original one in an attempt to 'self-repair'.
Not sure if it worked out exactly how I pictured but the nous eval's not overall much worse than the original at least.
(Took I think about 8 hours for, I want to say, ~80 million tokens on one RTX 3090?)
Thought about doing LORA first but I couldn't get peft_layers_to_transform working on axolotl and decided to go straight to LISA.
It's probably good(?) for (random selection of layers) to get experience trying to work around (random thick doubled layers) in some kind of brain-exercise sense anyway.
Capabilities assessment vs original and upscaled version:
| Model | AGIEval | GPT4All | TruthfulQA | Bigbench | Average |
|---|---|---|---|---|---|
| cosmo-upscale-lisa | 22.76 | 49.69 | 39.49 | 28.7 | 35.16 |
| Model | AGIEval | GPT4All | TruthfulQA | Bigbench | Average |
|---|---|---|---|---|---|
| cosmo-1b | 22.97 | 52.01 | 38.02 | 28.73 | 35.43 |
| Model | AGIEval | GPT4All | TruthfulQA | Bigbench | Average |
|---|---|---|---|---|---|
| cosmo-upscale | 22.23 | 48.35 | 42.01 | 28.36 | 35.24 |
I'm not sure this helped.
See axolotl config
axolotl version: 0.4.0
base_model: Lambent/cosmo-upscale
model_type: LlamaForCausalLM
tokenizer_type: LlamaTokenizer
load_in_8bit: false
load_in_4bit: false
strict: false
datasets:
- path: HuggingFaceTB/cosmopedia-100k
type: completion
- path: Vezora/Tested-22k-Python-Alpaca
type: alpaca
dataset_prepared_path:
val_set_size: 0.05
output_dir: ./lisa-out
sequence_len: 2048
sample_packing: true
pad_to_sequence_len: true
adapter:
lora_model_dir:
lora_r:
lora_alpha:
lora_dropout:
lora_target_linear:
lora_fan_in_fan_out:
lisa_n_layers: 4
lisa_step_interval: 10
lisa_layers_attribute: model.layers
wandb_project: cosmouplisa
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:
gradient_accumulation_steps: 4
micro_batch_size: 2
num_epochs: 1
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 0.0002
train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: false
gradient_checkpointing: true
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true
warmup_steps: 10
evals_per_epoch: 4
saves_per_epoch: 1
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:
lisa-out
This model is a fine-tuned version of Lambent/cosmo-upscale on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.0353
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0002
- train_batch_size: 2
- eval_batch_size: 2
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 10
- num_epochs: 1
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.4298 | 0.0 | 1 | 1.4591 |
| 1.1229 | 0.25 | 1480 | 1.0594 |
| 1.0711 | 0.5 | 2960 | 1.0418 |
| 1.0511 | 0.75 | 4440 | 1.0353 |
Framework versions
- Transformers 4.40.0.dev0
- Pytorch 2.1.2+cu118
- Datasets 2.18.0
- Tokenizers 0.15.0
- Downloads last month
- 2