unitorch.models.llama¤
LlamaProcessor¤
Bases: HfTextClassificationProcessor
, HfTextGenerationProcessor
Initialize the LlamaProcessor.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
vocab_file |
str
|
Path to the vocabulary file. |
required |
max_seq_length |
int
|
Maximum sequence length for text classification. Defaults to 128. |
128
|
max_gen_seq_length |
int
|
Maximum sequence length for text generation. Defaults to 48. |
48
|
Source code in src/unitorch/models/llama/processing.py
22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 |
|
classification ¤
classification(
text: str,
text_pair: Optional[str] = None,
max_seq_length: Optional[int] = None,
)
Process text for classification.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
text |
str
|
Input text. |
required |
text_pair |
str
|
Input text pair. Defaults to None. |
None
|
max_seq_length |
int
|
Maximum sequence length. Defaults to None. |
None
|
Returns:
Name | Type | Description |
---|---|---|
GenericOutputs |
Processed input_ids and attention_mask tensors. |
Source code in src/unitorch/models/llama/processing.py
65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 |
|
generation ¤
generation(
text: str,
text_pair: str,
max_seq_length: Optional[int] = None,
max_gen_seq_length: Optional[int] = None,
)
Process text for generation.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
text |
str
|
Input text. |
required |
text_pair |
str
|
Input text pair. |
required |
max_seq_length |
int
|
Maximum sequence length. Defaults to None. |
None
|
max_gen_seq_length |
int
|
Maximum generation sequence length. Defaults to None. |
None
|
Returns:
Name | Type | Description |
---|---|---|
GenericOutputs |
Processed input_ids, attention_mask, input_ids_label, and attention_mask_label tensors. |
Source code in src/unitorch/models/llama/processing.py
216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 |
|
generation_inputs ¤
generation_inputs(
text: str, max_seq_length: Optional[int] = None
)
Process text for generation inputs.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
text |
str
|
Input text. |
required |
max_seq_length |
int
|
Maximum sequence length. Defaults to None. |
None
|
Returns:
Name | Type | Description |
---|---|---|
GenericOutputs |
Processed input_ids tensor. |
Source code in src/unitorch/models/llama/processing.py
108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 |
|
generation_labels ¤
generation_labels(
text: str, max_gen_seq_length: Optional[int] = None
)
Process text for generation labels.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
text |
str
|
Input text. |
required |
max_gen_seq_length |
int
|
Maximum generation sequence length. Defaults to None. |
None
|
Returns:
Name | Type | Description |
---|---|---|
GenericOutputs |
Processed input_ids and attention_mask tensors. |
Source code in src/unitorch/models/llama/processing.py
179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 |
|
instruction_generation_inputs ¤
instruction_generation_inputs(
instruction: str,
input: str,
max_seq_length: Optional[int] = None,
) -> GenericOutputs
Preprocesses text as generation inputs.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
text |
str
|
The input text for generation. |
required |
max_seq_length |
Optional[int]
|
The maximum sequence length. Defaults to None. |
None
|
Returns:
Name | Type | Description |
---|---|---|
GenericOutputs |
GenericOutputs
|
The processed input IDs tensor. |
Source code in src/unitorch/models/llama/processing.py
148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 |
|
LlamaForClassification¤
Bases: GenericModel
, QuantizationMixin
, PeftWeightLoaderMixin
Llama model for classification tasks.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
config_path |
str
|
Path to the model configuration file. |
required |
num_classes |
int
|
Number of classes for classification. Defaults to 1. |
1
|
hidden_dropout_prob |
float
|
Dropout probability for hidden layers. Defaults to 0.1. |
0.1
|
gradient_checkpointing |
bool
|
Whether to use gradient checkpointing. Defaults to False. |
False
|
Source code in src/unitorch/models/llama/modeling.py
27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 |
|
forward ¤
forward(
input_ids: Tensor,
attention_mask: Optional[Tensor] = None,
position_ids: Optional[Tensor] = None,
)
Forward pass of the classification model.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
input_ids |
Tensor
|
Input tensor of shape (batch_size, sequence_length). |
required |
attention_mask |
Tensor
|
Attention mask tensor of shape (batch_size, sequence_length). Defaults to None. |
None
|
position_ids |
Tensor
|
Position IDs tensor of shape (batch_size, sequence_length). Defaults to None. |
None
|
Returns:
Type | Description |
---|---|
torch Output logits.Tensor: tensor of shape (batch_size, num_classes). |
Source code in src/unitorch/models/llama/modeling.py
56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 |
|
LlamaForGeneration¤
Bases: GenericModel
, QuantizationMixin
, PeftWeightLoaderMixin
Llama model for text generation tasks.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
config_path |
str
|
Path to the model configuration file. |
required |
gradient_checkpointing |
bool
|
Whether to use gradient checkpointing. Defaults to False. |
False
|
Source code in src/unitorch/models/llama/modeling.py
91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 |
|
forward ¤
forward(
input_ids: Tensor,
attention_mask: Optional[Tensor] = None,
position_ids: Optional[Tensor] = None,
)
Forward pass of the generation model.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
input_ids |
Tensor
|
Input tensor of shape (batch_size, sequence_length). Defaults to None. |
required |
attention_mask |
Tensor
|
Attention mask tensor of shape (batch_size, sequence_length). Defaults to None. |
None
|
position_ids |
Tensor
|
Position IDs tensor of shape (batch_size, sequence_length). Defaults to None. |
None
|
Returns:
Type | Description |
---|---|
torch Output logits.Tensor: tensor of shape (batch_size, sequence_length, vocab_size). |
Source code in src/unitorch/models/llama/modeling.py
114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 |
|
generate ¤
generate(
input_ids: Tensor,
num_beams: Optional[int] = 5,
decoder_start_token_id: Optional[int] = 1,
decoder_end_token_id: Optional[
Union[int, List[int]]
] = 2,
num_return_sequences: Optional[int] = 1,
min_gen_seq_length: Optional[int] = 0,
max_gen_seq_length: Optional[int] = 48,
repetition_penalty: Optional[float] = 1.0,
no_repeat_ngram_size: Optional[int] = 0,
early_stopping: Optional[bool] = True,
length_penalty: Optional[float] = 1.0,
num_beam_groups: Optional[int] = 1,
diversity_penalty: Optional[float] = 0.0,
do_sample: Optional[bool] = False,
temperature: Optional[float] = 1.0,
top_k: Optional[int] = 50,
top_p: Optional[float] = 1.0,
)
Generate text using the generation model.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
input_ids |
Tensor
|
Input tensor of shape (batch_size, sequence_length). |
required |
num_beams |
int
|
Number of beams for beam search. Defaults to 5. |
5
|
decoder_start_token_id |
int
|
The ID of the decoder start token. Defaults to 2. |
1
|
decoder_end_token_id |
int or List[int]
|
The ID(s) of the decoder end token(s). Defaults to 2. |
2
|
num_return_sequences |
int
|
Number of generated sequences to return. Defaults to 1. |
1
|
min_gen_seq_length |
int
|
Minimum length of generated sequences. Defaults to 0. |
0
|
max_gen_seq_length |
int
|
Maximum length of generated sequences. Defaults to 48. |
48
|
repetition_penalty |
float
|
Penalty for repeated tokens. Defaults to 1.0. |
1.0
|
no_repeat_ngram_size |
int
|
Size of n-grams to avoid repeating. Defaults to 0. |
0
|
early_stopping |
bool
|
Whether to stop generation early. Defaults to True. |
True
|
length_penalty |
float
|
Penalty for longer sequences. Defaults to 1.0. |
1.0
|
num_beam_groups |
int
|
Number of beam groups for diverse beam search. Defaults to 1. |
1
|
diversity_penalty |
float
|
Penalty for diverse sequences in diverse beam search. Defaults to 0.0. |
0.0
|
do_sample |
bool
|
Whether to use sampling for generation. Defaults to False. |
False
|
temperature |
float
|
Sampling temperature. Defaults to 1.0. |
1.0
|
top_k |
int
|
Top-k value for sampling. Defaults to 50. |
50
|
top_p |
float
|
Top-p value for sampling. Defaults to 1.0. |
1.0
|
Returns:
Name | Type | Description |
---|---|---|
GenericOutputs |
Generated sequences and their scores. |
Source code in src/unitorch/models/llama/modeling.py
140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 |
|