Jeopardy bot! (#17)
* support for jeopardy dataset * commit the final config for jeopardy bot
This commit is contained in:
58
configs/llama_7B_jeopardy.yml
Normal file
58
configs/llama_7B_jeopardy.yml
Normal file
@@ -0,0 +1,58 @@
|
|||||||
|
base_model: huggyllama/llama-7b
|
||||||
|
base_model_config: huggyllama/llama-7b
|
||||||
|
model_type: LlamaForCausalLM
|
||||||
|
tokenizer_type: LlamaTokenizer
|
||||||
|
load_in_8bit: false
|
||||||
|
datasets:
|
||||||
|
- path: openaccess-ai-collective/jeopardy
|
||||||
|
type: jeopardy
|
||||||
|
dataset_prepared_path: last_run_prepared
|
||||||
|
val_set_size: 0.01
|
||||||
|
adapter:
|
||||||
|
lora_model_dir:
|
||||||
|
sequence_len: 2048
|
||||||
|
max_packed_sequence_len: 2048
|
||||||
|
lora_r: 8
|
||||||
|
lora_alpha: 16
|
||||||
|
lora_dropout: 0.05
|
||||||
|
lora_target_modules:
|
||||||
|
- q_proj
|
||||||
|
- v_proj
|
||||||
|
lora_fan_in_fan_out: false
|
||||||
|
wandb_project: jeopardy-bot-7b
|
||||||
|
wandb_watch:
|
||||||
|
wandb_run_id:
|
||||||
|
wandb_log_model: checkpoint
|
||||||
|
output_dir: ./jeopardy-bot-7b
|
||||||
|
batch_size: 4
|
||||||
|
micro_batch_size: 1
|
||||||
|
num_epochs: 2
|
||||||
|
optimizer: adamw_bnb_8bit
|
||||||
|
torchdistx_path:
|
||||||
|
lr_scheduler: cosine
|
||||||
|
learning_rate: 0.0000002
|
||||||
|
train_on_inputs: false
|
||||||
|
group_by_length: false
|
||||||
|
bf16: true
|
||||||
|
tf32: true
|
||||||
|
early_stopping_patience:
|
||||||
|
resume_from_checkpoint:
|
||||||
|
local_rank:
|
||||||
|
logging_steps: 5
|
||||||
|
xformers_attention: true
|
||||||
|
flash_attention:
|
||||||
|
gptq_groupsize:
|
||||||
|
gptq_model_v1:
|
||||||
|
warmup_steps: 20
|
||||||
|
eval_steps: 110
|
||||||
|
save_steps: 660
|
||||||
|
debug:
|
||||||
|
deepspeed:
|
||||||
|
weight_decay: 0.0001
|
||||||
|
fsdp:
|
||||||
|
fsdp_config:
|
||||||
|
special_tokens:
|
||||||
|
pad_token: "[PAD]"
|
||||||
|
bos_token: "<s>"
|
||||||
|
eos_token: "</s>"
|
||||||
|
unk_token: "<unk>"
|
||||||
@@ -89,6 +89,15 @@ class AlpacaPromptTokenizingStrategy(InstructionPromptTokenizingStrategy):
|
|||||||
)
|
)
|
||||||
|
|
||||||
|
|
||||||
|
class JeopardyPromptTokenizingStrategy(InstructionPromptTokenizingStrategy):
|
||||||
|
def parse_instruction_fields(self, prompt) -> (str, str, str):
|
||||||
|
return (
|
||||||
|
prompt["question"],
|
||||||
|
prompt["category"],
|
||||||
|
"what is " + prompt["answer"],
|
||||||
|
)
|
||||||
|
|
||||||
|
|
||||||
class OpenAssistantPromptTokenizingStrategy(InstructionPromptTokenizingStrategy):
|
class OpenAssistantPromptTokenizingStrategy(InstructionPromptTokenizingStrategy):
|
||||||
def parse_instruction_fields(self, prompt) -> (str, str, str):
|
def parse_instruction_fields(self, prompt) -> (str, str, str):
|
||||||
return (
|
return (
|
||||||
|
|||||||
@@ -31,6 +31,10 @@ class AlpacaPrompter:
|
|||||||
return output.split(self.response_split)[1].strip()
|
return output.split(self.response_split)[1].strip()
|
||||||
|
|
||||||
|
|
||||||
|
class JeopardyPrompter(AlpacaPrompter):
|
||||||
|
prompt_input = "Below is a Jeopardy clue paired with input providing the category of the clue. Write a concise response that best answers tbe clue given the category.\n\n### Instruction:\n{instruction}\n\n### Input:\n{input}\n\n### Response:\n"
|
||||||
|
|
||||||
|
|
||||||
class GPTeacherPrompter(AlpacaPrompter):
|
class GPTeacherPrompter(AlpacaPrompter):
|
||||||
...
|
...
|
||||||
|
|
||||||
|
|||||||
@@ -11,13 +11,13 @@ from axolotl.prompt_tokenizers import (
|
|||||||
GPTeacherPromptTokenizingStrategy,
|
GPTeacherPromptTokenizingStrategy,
|
||||||
OpenAssistantPromptTokenizingStrategy,
|
OpenAssistantPromptTokenizingStrategy,
|
||||||
AlpacaReflectionPTStrategy,
|
AlpacaReflectionPTStrategy,
|
||||||
ShareGPTPromptTokenizingStrategy,
|
ShareGPTPromptTokenizingStrategy, JeopardyPromptTokenizingStrategy,
|
||||||
)
|
)
|
||||||
from axolotl.prompters import (
|
from axolotl.prompters import (
|
||||||
AlpacaPrompter,
|
AlpacaPrompter,
|
||||||
GPTeacherPrompter,
|
GPTeacherPrompter,
|
||||||
ReflectAlpacaPrompter,
|
ReflectAlpacaPrompter,
|
||||||
ShareGPTPrompter,
|
ShareGPTPrompter, JeopardyPrompter,
|
||||||
)
|
)
|
||||||
|
|
||||||
|
|
||||||
@@ -82,6 +82,12 @@ def load_prepare_datasets(tokenizer, cfg, default_dataset_prepared_path):
|
|||||||
)
|
)
|
||||||
ds_wrapper = TokenizedPromptDataset(ds_strategy, ds["train"])
|
ds_wrapper = TokenizedPromptDataset(ds_strategy, ds["train"])
|
||||||
datasets.append(ds_wrapper)
|
datasets.append(ds_wrapper)
|
||||||
|
if d.type == "jeopardy":
|
||||||
|
ds_strategy = JeopardyPromptTokenizingStrategy(
|
||||||
|
JeopardyPrompter(), tokenizer, cfg.train_on_inputs, cfg.sequence_len
|
||||||
|
)
|
||||||
|
ds_wrapper = TokenizedPromptDataset(ds_strategy, ds["train"])
|
||||||
|
datasets.append(ds_wrapper)
|
||||||
elif d.type == "oasst":
|
elif d.type == "oasst":
|
||||||
ds_strategy = OpenAssistantPromptTokenizingStrategy(
|
ds_strategy = OpenAssistantPromptTokenizingStrategy(
|
||||||
AlpacaPrompter(), tokenizer, cfg.train_on_inputs, cfg.sequence_len
|
AlpacaPrompter(), tokenizer, cfg.train_on_inputs, cfg.sequence_len
|
||||||
|
|||||||
Reference in New Issue
Block a user