Skip to content

Commit

Permalink
Merge pull request #5039 from oobabooga/dev
Browse files Browse the repository at this point in the history 
Merge dev branch
  • Loading branch information
@oobabooga
oobabooga authored Dec 21, 2023
2 parents 11288d1 + c1b99f4 commit 4b25acf
Show file tree
Hide file tree
Showing 4 changed files with 137 additions and 293 deletions.
2 changes: 1 addition & 1 deletion README.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -237,7 +237,7 @@ List of command-line flags
| `--no_use_fast` | Set use_fast=False while loading the tokenizer (it's True by default). Use this if you have any problems related to use_fast. |
| `--use_flash_attention_2` | Set use_flash_attention_2=True while loading the model. |

#### Accelerate 4-bit
#### bitsandbytes 4-bit

⚠️ Requires minimum compute of 7.0 on Windows at the moment.

Expand Down
201 changes: 13 additions & 188 deletions modules/loaders.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -168,8 +168,9 @@
]
})

loaders_samplers = {
'Transformers': {

def transformers_samplers():
return {
'temperature',
'temperature_last',
'top_p',
Expand Down Expand Up @@ -205,7 +206,16 @@
'add_bos_token',
'skip_special_tokens',
'auto_max_new_tokens',
},
}


loaders_samplers = {
'Transformers': transformers_samplers(),
'AutoGPTQ': transformers_samplers(),
'GPTQ-for-LLaMa': transformers_samplers(),
'AutoAWQ': transformers_samplers(),
'QuIP#': transformers_samplers(),
'HQQ': transformers_samplers(),
'ExLlama_HF': {
'temperature',
'temperature_last',
Expand Down Expand Up @@ -306,80 +316,6 @@
'skip_special_tokens',
'auto_max_new_tokens',
},
'AutoGPTQ': {
'temperature',
'temperature_last',
'top_p',
'min_p',
'top_k',
'typical_p',
'epsilon_cutoff',
'eta_cutoff',
'tfs',
'top_a',
'repetition_penalty',
'presence_penalty',
'frequency_penalty',
'repetition_penalty_range',
'encoder_repetition_penalty',
'no_repeat_ngram_size',
'min_length',
'seed',
'do_sample',
'penalty_alpha',
'num_beams',
'length_penalty',
'early_stopping',
'mirostat_mode',
'mirostat_tau',
'mirostat_eta',
'grammar_file_row',
'grammar_string',
'guidance_scale',
'negative_prompt',
'ban_eos_token',
'custom_token_bans',
'add_bos_token',
'skip_special_tokens',
'auto_max_new_tokens',
},
'GPTQ-for-LLaMa': {
'temperature',
'temperature_last',
'top_p',
'min_p',
'top_k',
'typical_p',
'epsilon_cutoff',
'eta_cutoff',
'tfs',
'top_a',
'repetition_penalty',
'presence_penalty',
'frequency_penalty',
'repetition_penalty_range',
'encoder_repetition_penalty',
'no_repeat_ngram_size',
'min_length',
'seed',
'do_sample',
'penalty_alpha',
'num_beams',
'length_penalty',
'early_stopping',
'mirostat_mode',
'mirostat_tau',
'mirostat_eta',
'grammar_file_row',
'grammar_string',
'guidance_scale',
'negative_prompt',
'ban_eos_token',
'custom_token_bans',
'add_bos_token',
'skip_special_tokens',
'auto_max_new_tokens',
},
'llama.cpp': {
'temperature',
'top_p',
Expand Down Expand Up @@ -439,117 +375,6 @@
'repetition_penalty',
'repetition_penalty_range',
},
'AutoAWQ': {
'temperature',
'temperature_last',
'top_p',
'min_p',
'top_k',
'typical_p',
'epsilon_cutoff',
'eta_cutoff',
'tfs',
'top_a',
'repetition_penalty',
'presence_penalty',
'frequency_penalty',
'repetition_penalty_range',
'encoder_repetition_penalty',
'no_repeat_ngram_size',
'min_length',
'seed',
'do_sample',
'penalty_alpha',
'num_beams',
'length_penalty',
'early_stopping',
'mirostat_mode',
'mirostat_tau',
'mirostat_eta',
'grammar_file_row',
'grammar_string',
'guidance_scale',
'negative_prompt',
'ban_eos_token',
'custom_token_bans',
'add_bos_token',
'skip_special_tokens',
'auto_max_new_tokens',
},
'QuIP#': {
'temperature',
'temperature_last',
'top_p',
'min_p',
'top_k',
'typical_p',
'epsilon_cutoff',
'eta_cutoff',
'tfs',
'top_a',
'repetition_penalty',
'presence_penalty',
'frequency_penalty',
'repetition_penalty_range',
'encoder_repetition_penalty',
'no_repeat_ngram_size',
'min_length',
'seed',
'do_sample',
'penalty_alpha',
'num_beams',
'length_penalty',
'early_stopping',
'mirostat_mode',
'mirostat_tau',
'mirostat_eta',
'grammar_file_row',
'grammar_string',
'guidance_scale',
'negative_prompt',
'ban_eos_token',
'custom_token_bans',
'add_bos_token',
'skip_special_tokens',
'auto_max_new_tokens',
},
'HQQ': {
'temperature',
'temperature_last',
'top_p',
'min_p',
'top_k',
'typical_p',
'epsilon_cutoff',
'eta_cutoff',
'tfs',
'top_a',
'repetition_penalty',
'presence_penalty',
'frequency_penalty',
'repetition_penalty_range',
'encoder_repetition_penalty',
'no_repeat_ngram_size',
'min_length',
'seed',
'do_sample',
'penalty_alpha',
'num_beams',
'length_penalty',
'early_stopping',
'mirostat_mode',
'mirostat_tau',
'mirostat_eta',
'grammar_file_row',
'grammar_string',
'guidance_scale',
'negative_prompt',
'ban_eos_token',
'custom_token_bans',
'add_bos_token',
'skip_special_tokens',
'auto_max_new_tokens',
},
}

loaders_model_types = {
Expand Down
Loading

0 comments on commit 4b25acf

Please sign in to comment.