|
--- |
|
base_model: s3nh/phi-1_5_dolly_instruction_polish |
|
inference: false |
|
language: |
|
- pl |
|
- en |
|
library_name: transformers |
|
license: openrail |
|
model_creator: s3nh |
|
model_name: phi-1_5_dolly_instruction_polish |
|
pipeline_tag: text-generation |
|
quantized_by: afrideva |
|
tags: |
|
- gguf |
|
- ggml |
|
- quantized |
|
- q2_k |
|
- q3_k_m |
|
- q4_k_m |
|
- q5_k_m |
|
- q6_k |
|
- q8_0 |
|
--- |
|
# s3nh/phi-1_5_dolly_instruction_polish-GGUF |
|
|
|
Quantized GGUF model files for [phi-1_5_dolly_instruction_polish](https://huggingface.co/s3nh/phi-1_5_dolly_instruction_polish) from [s3nh](https://huggingface.co/s3nh) |
|
|
|
|
|
| Name | Quant method | Size | |
|
| ---- | ---- | ---- | |
|
| [phi-1_5_dolly_instruction_polish.fp16.gguf](https://huggingface.co/afrideva/phi-1_5_dolly_instruction_polish-GGUF/resolve/main/phi-1_5_dolly_instruction_polish.fp16.gguf) | fp16 | 2.84 GB | |
|
| [phi-1_5_dolly_instruction_polish.q2_k.gguf](https://huggingface.co/afrideva/phi-1_5_dolly_instruction_polish-GGUF/resolve/main/phi-1_5_dolly_instruction_polish.q2_k.gguf) | q2_k | 612.98 MB | |
|
| [phi-1_5_dolly_instruction_polish.q3_k_m.gguf](https://huggingface.co/afrideva/phi-1_5_dolly_instruction_polish-GGUF/resolve/main/phi-1_5_dolly_instruction_polish.q3_k_m.gguf) | q3_k_m | 765.45 MB | |
|
| [phi-1_5_dolly_instruction_polish.q4_k_m.gguf](https://huggingface.co/afrideva/phi-1_5_dolly_instruction_polish-GGUF/resolve/main/phi-1_5_dolly_instruction_polish.q4_k_m.gguf) | q4_k_m | 918.31 MB | |
|
| [phi-1_5_dolly_instruction_polish.q5_k_m.gguf](https://huggingface.co/afrideva/phi-1_5_dolly_instruction_polish-GGUF/resolve/main/phi-1_5_dolly_instruction_polish.q5_k_m.gguf) | q5_k_m | 1.06 GB | |
|
| [phi-1_5_dolly_instruction_polish.q6_k.gguf](https://huggingface.co/afrideva/phi-1_5_dolly_instruction_polish-GGUF/resolve/main/phi-1_5_dolly_instruction_polish.q6_k.gguf) | q6_k | 1.17 GB | |
|
| [phi-1_5_dolly_instruction_polish.q8_0.gguf](https://huggingface.co/afrideva/phi-1_5_dolly_instruction_polish-GGUF/resolve/main/phi-1_5_dolly_instruction_polish.q8_0.gguf) | q8_0 | 1.51 GB | |
|
|
|
|
|
|
|
## Original Model Card: |
|
microsoft/phi_1.5 finetuned on s3nh/dolly_instruction_polish. |
|
|
|
Finetuned with QLora, provided version is adapter merged with base model. |
|
Load in 4 bit, sequence length set to 1024. |
|
|
|
|
|
axolotl config |
|
``` |
|
base_model: microsoft/phi-2 |
|
model_type: AutoModelForCausalLM |
|
tokenizer_type: AutoTokenizer |
|
is_llama_derived_model: false |
|
trust_remote_code: true |
|
|
|
load_in_8bit: false |
|
load_in_4bit: true |
|
strict: false |
|
|
|
datasets: |
|
- path: s3nh/alpaca-dolly-instruction-only-polish |
|
type: alpaca |
|
|
|
dataset_prepared_path: |
|
val_set_size: 0.05 |
|
output_dir: ./phi-2-sft-out |
|
|
|
sequence_len: 1024 |
|
sample_packing: false # not CURRENTLY compatible with LoRAs |
|
pad_to_sequence_len: |
|
|
|
adapter: qlora |
|
lora_model_dir: |
|
lora_r: 64 |
|
lora_alpha: 32 |
|
lora_dropout: 0.05 |
|
lora_target_linear: true |
|
lora_fan_in_fan_out: |
|
|
|
wandb_project: |
|
wandb_entity: |
|
wandb_watch: |
|
wandb_name: |
|
wandb_log_model: |
|
|
|
gradient_accumulation_steps: 1 |
|
micro_batch_size: 1 |
|
num_epochs: 4 |
|
optimizer: adamw_torch |
|
adam_beta2: 0.95 |
|
adam_epsilon: 0.00001 |
|
max_grad_norm: 1.0 |
|
lr_scheduler: cosine |
|
learning_rate: 0.000003 |
|
|
|
train_on_inputs: false |
|
group_by_length: true |
|
bf16: true |
|
fp16: false |
|
tf32: true |
|
|
|
gradient_checkpointing: |
|
early_stopping_patience: |
|
resume_from_checkpoint: false |
|
local_rank: |
|
logging_steps: 100 |
|
xformers_attention: |
|
flash_attention: true |
|
|
|
warmup_steps: 10 |
|
evals_per_epoch: 4 |
|
saves_per_epoch: |
|
save_strategy: steps |
|
save_steps: 5000 |
|
debug: |
|
deepspeed: |
|
weight_decay: 0.1 |
|
fsdp: |
|
fsdp_config: |
|
resize_token_embeddings_to_32x: true |
|
special_tokens: |
|
bos_token: "<|endoftext|>" |
|
eos_token: "<|endoftext|>" |
|
unk_token: "<|endoftext|>" |
|
pad_token: "<|endoftext|>" |
|
``` |
