RASMUS commited on
Commit
71ff162
1 Parent(s): 45796ea

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +221 -197
README.md CHANGED
@@ -1,201 +1,225 @@
1
  ---
2
  library_name: transformers
3
- tags: []
 
 
 
 
 
4
  ---
5
 
6
- # Model Card for Model ID
7
-
8
- <!-- Provide a quick summary of what the model is/does. -->
9
-
10
-
11
-
12
- ## Model Details
13
-
14
- ### Model Description
15
-
16
- <!-- Provide a longer summary of what this model is. -->
17
-
18
- This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
19
-
20
- - **Developed by:** [More Information Needed]
21
- - **Funded by [optional]:** [More Information Needed]
22
- - **Shared by [optional]:** [More Information Needed]
23
- - **Model type:** [More Information Needed]
24
- - **Language(s) (NLP):** [More Information Needed]
25
- - **License:** [More Information Needed]
26
- - **Finetuned from model [optional]:** [More Information Needed]
27
-
28
- ### Model Sources [optional]
29
-
30
- <!-- Provide the basic links for the model. -->
31
-
32
- - **Repository:** [More Information Needed]
33
- - **Paper [optional]:** [More Information Needed]
34
- - **Demo [optional]:** [More Information Needed]
35
-
36
- ## Uses
37
-
38
- <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
39
-
40
- ### Direct Use
41
-
42
- <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
43
-
44
- [More Information Needed]
45
-
46
- ### Downstream Use [optional]
47
-
48
- <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
49
-
50
- [More Information Needed]
51
-
52
- ### Out-of-Scope Use
53
-
54
- <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
55
-
56
- [More Information Needed]
57
-
58
- ## Bias, Risks, and Limitations
59
-
60
- <!-- This section is meant to convey both technical and sociotechnical limitations. -->
61
-
62
- [More Information Needed]
63
-
64
- ### Recommendations
65
-
66
- <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
67
-
68
- Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
69
-
70
- ## How to Get Started with the Model
71
-
72
- Use the code below to get started with the model.
73
-
74
- [More Information Needed]
75
-
76
- ## Training Details
77
-
78
- ### Training Data
79
-
80
- <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
81
-
82
- [More Information Needed]
83
-
84
- ### Training Procedure
85
-
86
- <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
87
-
88
- #### Preprocessing [optional]
89
-
90
- [More Information Needed]
91
-
92
-
93
- #### Training Hyperparameters
94
-
95
- - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
96
-
97
- #### Speeds, Sizes, Times [optional]
98
-
99
- <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
100
-
101
- [More Information Needed]
102
-
103
- ## Evaluation
104
-
105
- <!-- This section describes the evaluation protocols and provides the results. -->
106
-
107
- ### Testing Data, Factors & Metrics
108
-
109
- #### Testing Data
110
-
111
- <!-- This should link to a Dataset Card if possible. -->
112
-
113
- [More Information Needed]
114
-
115
- #### Factors
116
-
117
- <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
118
-
119
- [More Information Needed]
120
-
121
- #### Metrics
122
-
123
- <!-- These are the evaluation metrics being used, ideally with a description of why. -->
124
-
125
- [More Information Needed]
126
-
127
- ### Results
128
-
129
- [More Information Needed]
130
-
131
- #### Summary
132
-
133
-
134
-
135
- ## Model Examination [optional]
136
-
137
- <!-- Relevant interpretability work for the model goes here -->
138
-
139
- [More Information Needed]
140
-
141
- ## Environmental Impact
142
-
143
- <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
144
-
145
- Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
146
-
147
- - **Hardware Type:** [More Information Needed]
148
- - **Hours used:** [More Information Needed]
149
- - **Cloud Provider:** [More Information Needed]
150
- - **Compute Region:** [More Information Needed]
151
- - **Carbon Emitted:** [More Information Needed]
152
-
153
- ## Technical Specifications [optional]
154
-
155
- ### Model Architecture and Objective
156
-
157
- [More Information Needed]
158
-
159
- ### Compute Infrastructure
160
-
161
- [More Information Needed]
162
-
163
- #### Hardware
164
-
165
- [More Information Needed]
166
-
167
- #### Software
168
-
169
- [More Information Needed]
170
-
171
- ## Citation [optional]
172
-
173
- <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
174
-
175
- **BibTeX:**
176
-
177
- [More Information Needed]
178
-
179
- **APA:**
180
-
181
- [More Information Needed]
182
-
183
- ## Glossary [optional]
184
-
185
- <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
186
-
187
- [More Information Needed]
188
-
189
- ## More Information [optional]
190
-
191
- [More Information Needed]
192
-
193
- ## Model Card Authors [optional]
194
-
195
- [More Information Needed]
196
-
197
- ## Model Card Contact
198
-
199
- [More Information Needed]
200
-
201
-
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  library_name: transformers
3
+ license: apache-2.0
4
+ tags:
5
+ - finnish
6
+ - llama
7
+ inference: true
8
+ pipeline_tag: text-generation
9
  ---
10
 
11
+ # Llama-7b-instruct-v0.2 for Finnish
12
+
13
+
14
+ - This is 0.2 version release of our Instruct finetuned model from https://huggingface.co/Finnish-NLP/llama-7b-finnish
15
+ - Model was trained for 3 epochs using 21946 samples and for this release we chose checkpoint at 8000 steps.
16
+ - Future DPO/SFT+DPO variants are in the pipeline. Also we are investigating and testing different merging techiques
17
+
18
+ For finetuning we try to select well known and widely used dataset and then filter/translate those with multiple methods:
19
+ For this version we used a mix 21946 samples in total from the the following datasets:
20
+ - LIMA from https://github.com/TurkuNLP/finnish-instructions
21
+ - Dolly from https://github.com/TurkuNLP/finnish-instructions
22
+ - OASST from https://github.com/TurkuNLP/finnish-instructions
23
+ - Ultrachat https://huggingface.co/datasets/HuggingFaceH4/ultrafeedback_binarized/viewer/default/train_sft translated with deepl
24
+ - facebook/belebele Finnish subset
25
+ - google/boolq translated with deepl
26
+ - LDJnr/Capybara translated with deepl
27
+ - allenai/ai2_arc translated with deepl
28
+
29
+
30
+ ### How to use
31
+
32
+ Here is an example of using this model with Unsloth with some generation arguments you can modify:
33
+
34
+ ```python
35
+ import torch
36
+ from unsloth import FastLlamaModel
37
+
38
+ max_seq_length = 2048
39
+ dtype = None # None for auto detection. Float16 for Tesla T4, V100, Bfloat16 for Ampere+
40
+ load_in_4bit = True # Use 4bit quantization to reduce memory usage. Can be False.
41
+
42
+
43
+ use_unsloth = True
44
+ # use_transformers = True
45
+
46
+ # LOADING MODEL USIINIG TRANSFORMERS assumes at least 16GB of memory. Tested with this configuration
47
+ # If you have less memory use load_in_4bit or load_in_8_bit as needed
48
+ if use_transformers:
49
+ major_version, minor_version = torch.cuda.get_device_capability()
50
+ model = AutoModelForCausalLM.from_pretrained("Finnish-NLP/llama-7b-finnish-instruct-v0.2", device_map='cuda:0', torch_dtype = torch.bfloat16 if major_version >=8 else torch.float16)
51
+ tokenizer = AutoTokenizer.from_pretrained("Finnish-NLP/llama-7b-finnish-instruct-v0.2")
52
+
53
+ # USING UNSLOTH, tested with load_in_4bit
54
+ if use_unsloth:
55
+ model, tokenizer = FastLlamaModel.from_pretrained(
56
+ model_name = "Finnish-NLP/llama-7b-finnish-instruct-v0.2"
57
+ max_seq_length = max_seq_length,
58
+ dtype = dtype,
59
+ load_in_4bit = load_in_4bit
60
+ )
61
+
62
+ alpaca_prompt = """<|alku|> Olet tekoälyavustaja. Seuraavaksi saat kysymyksen tai tehtävän. Kirjoita vastaus parhaasi mukaan siten että se täyttää kysymyksen tai tehtävän vaatimukset.
63
+ <|ihminen|> Kysymys/Tehtävä:
64
+ {}
65
+ <|avustaja|> Vastauksesi:
66
+ """
67
+
68
+ sample_questions = ["Ketkä ovat Aku Ankan luona asuvat kolme ankanpoikaa?",\
69
+ "Mikä on Suomen korkein tunturi?",\
70
+ "Suomi soti Neuvostoliittoa vastaan talvisodan 1939-1940. Kuinka monta päivää sota kesti?",\
71
+ "Luettele viisi yleistä Suomessa yleisesti käytettyä pojan nimeä. Nimet:",\
72
+ "Luettele lyhyt, maksimissaan 50 sanan mittainen runo Suomesta. Runo:",\
73
+ ]
74
+
75
+ from transformers import GenerationConfig
76
+
77
+ generation_config = GenerationConfig(
78
+ pad_token_id=tokenizer.eos_token_id,
79
+ eos_token_id=tokenizer.convert_tokens_to_ids("<|loppu|>"),
80
+ )
81
+
82
+
83
+ for sample_question in sample_questions:
84
+
85
+ model.eval()
86
+
87
+ inputs = tokenizer(
88
+ [
89
+ alpaca_prompt.format(
90
+ sample_question, # instruction
91
+ )
92
+ ]*1, return_tensors = "pt").to("cuda")
93
+
94
+ with torch.no_grad():
95
+ generated_ids = model.generate(
96
+ input_ids=inputs["input_ids"],
97
+ attention_mask=inputs["attention_mask"],
98
+ generation_config=generation_config, **{
99
+ "temperature": 0.1,
100
+ "penalty_alpha": 0.6,
101
+ "top_k": 3,
102
+ "do_sample": True,
103
+ "repetition_penalty": 1.28,
104
+ "min_length": 10,
105
+ "max_new_tokens": 200
106
+ })
107
+
108
+ generated_text = tokenizer.batch_decode(generated_ids, skip_special_tokens=True, clean_up_tokenization_spaces=True)[0]
109
+ print(len(generated_ids[0]))
110
+ print("KYSYMYS:")
111
+ print(generated_text.split('<|avustaja|>')[0])
112
+ print("VASTAUS:")
113
+ print(generated_text.split('<|avustaja|> Vastauksesi:')[1])
114
+ print('##################################')
115
+
116
+ '''
117
+ -->
118
+ <s><|alku|> Olet tekoälyavustaja. Seuraavaksi saat kysymyksen tai tehtävän. Kirjoita vastaus parhaasi mukaan siten että se täyttää kysymyksen tai tehtävän vaatimukset.
119
+ <|ihminen|> Kysymys/Tehtävä:
120
+ Aku Ankan luona asuu kolme ankanpoikaa. Mitkä ovat heidän nimet?
121
+ VASTAUS:
122
+ Ankka Akun kanssa asuvat pojat ovat nimeltään Tupu, Hupu ja Lupu <|loppu|>
123
+ ##################################
124
+
125
+ KYSYMYS:
126
+ <s><|alku|> Olet tekoälyavustaja. Seuraavaksi saat kysymyksen tai tehtävän. Kirjoita vastaus parhaasi mukaan siten että se täyttää kysymyksen tai tehtävän vaatimukset.
127
+ <|ihminen|> Kysymys/Tehtävä:
128
+ Mikä on Suomen korkein tunturi?
129
+ VASTAUS:
130
+ Suomen korkein tunturihuippu on Haltitunturi (1 324 metriä). <|loppu|>
131
+ ##################################
132
+
133
+ KYSYMYS:
134
+ <s><|alku|> Olet tekoälyavustaja. Seuraavaksi saat kysymyksen tai tehtävän. Kirjoita vastaus parhaasi mukaan siten että se täyttää kysymyksen tai tehtävän vaatimukset.
135
+ <|ihminen|> Kysymys/Tehtävä:
136
+ Suomi soti Neuvostoliittoa vastaan talvisodan 1939-1940. Kuinka monta päivää sota kesti?
137
+ VASTAUS:
138
+ Talvisodan aikana Neuvostoliitto hyökkäsi Suomeen 30. marraskuuta ja 13. maaliskuuta välisenä aikana. Tämä tarkoittaa, että talvisota kesti 105 päivää. <|loppu|>
139
+ ##################################
140
+
141
+ KYSYMYS:
142
+ <s><|alku|> Olet tekoälyavustaja. Seuraavaksi saat kysymyksen tai tehtävän. Kirjoita vastaus parhaasi mukaan siten että se täyttää kysymyksen tai tehtävän vaatimukset.
143
+ <|ihminen|> Kysymys/Tehtävä:
144
+ Luettele viisi yleistä Suomessa yleisesti käytettyä pojan nimeä. Nimet:
145
+ VASTAUS:
146
+ Yleisiä suomalaisia poikien nimiä ovat Eino, Onni, Olavi, Väinö ja Ilmari. <|loppu|>
147
+ ##################################
148
+
149
+ KYSYMYS:
150
+ <s><|alku|> Olet tekoälyavustaja. Seuraavaksi saat kysymyksen tai tehtävän. Kirjoita vastaus parhaasi mukaan siten että se täyttää kysymyksen tai tehtävän vaatimukset.
151
+ <|ihminen|> Kysymys/Tehtävä:
152
+ Luettele lyhyt, maksimissaan 50 sanan mittainen runo Suomesta. Runo:
153
+ VASTAUS:
154
+ Olipa kerran kaunis maa,
155
+ jossa ihmiset elivät sopusoinnussa.
156
+ Se oli täynnä metsiä ja järviä,
157
+ ja siellä asui onnellisia ja ystävällisiä ihmisiä. <|loppu|>
158
+ ```
159
+
160
+ ### Limitations and bias
161
+
162
+ The training data used for this model contains a lot of content from the internet, which is far from neutral.
163
+ Therefore, the model can have biased predictions. This bias will also affect all fine-tuned versions of this model.
164
+ To reduce toxic content, the pretrained version of thiis model was trained with dataset filtered with a toxicity classifier but it cannot truly eliminate all toxic text.
165
+
166
+ ### Finetuning
167
+
168
+ Training was conducted on RTX 4080 using Unsloth framework https://github.com/unslothai/unsloth \
169
+ Training script is available in this repo.
170
+
171
+
172
+ ## Evaluation results
173
+
174
+ This model was evaluated using [FIN-bench by TurkuNLP](https://github.com/TurkuNLP/FIN-bench) with zero-shot setting, but \
175
+ the evaluation script had some problems running succesfully, so the results reported below should perhaps be viewed with some caution.
176
+
177
+ [llama-7b-finnish-instruct-v0.2](https://huggingface.co/Finnish-NLP/llama-7b-finnish-instruct-v0.2):
178
+
179
+ | Task |Version| Metric |Value | |Stderr|
180
+ |------------------------------------------------|------:|---------------------|-----:|---|-----:|
181
+ |bigbench_analogies | 0|multiple_choice_grade|0.5385|± |0.0439|
182
+ |bigbench_arithmetic_1_digit_addition | 0|multiple_choice_grade|0.3400|± |0.0476|
183
+ |bigbench_arithmetic_1_digit_division | 0|multiple_choice_grade|0.4783|± |0.1065|
184
+ |bigbench_arithmetic_1_digit_multiplication | 0|multiple_choice_grade|0.5200|± |0.0502|
185
+ |bigbench_arithmetic_1_digit_subtraction | 0|multiple_choice_grade|0.3400|± |0.0476|
186
+ |bigbench_arithmetic_2_digit_addition | 0|multiple_choice_grade|0.3200|± |0.0469|
187
+ |bigbench_arithmetic_2_digit_division | 0|multiple_choice_grade|0.3400|± |0.0476|
188
+ |bigbench_arithmetic_2_digit_multiplication | 0|multiple_choice_grade|0.2200|± |0.0416|
189
+ |bigbench_arithmetic_2_digit_subtraction | 0|multiple_choice_grade|0.2800|± |0.0451|
190
+ |bigbench_arithmetic_3_digit_addition | 0|multiple_choice_grade|0.3000|± |0.0461|
191
+ |bigbench_arithmetic_3_digit_division | 0|multiple_choice_grade|0.2500|± |0.0435|
192
+ |bigbench_arithmetic_3_digit_multiplication | 0|multiple_choice_grade|0.2200|± |0.0416|
193
+ |bigbench_arithmetic_3_digit_subtraction | 0|multiple_choice_grade|0.4000|± |0.0492|
194
+ |bigbench_arithmetic_4_digit_addition | 0|multiple_choice_grade|0.3500|± |0.0479|
195
+ |bigbench_arithmetic_4_digit_division | 0|multiple_choice_grade|0.2600|± |0.0441|
196
+ |bigbench_arithmetic_4_digit_multiplication | 0|multiple_choice_grade|0.2100|± |0.0409|
197
+ |bigbench_arithmetic_4_digit_subtraction | 0|multiple_choice_grade|0.4400|± |0.0499|
198
+ |bigbench_arithmetic_5_digit_addition | 0|multiple_choice_grade|0.4500|± |0.0500|
199
+ |bigbench_arithmetic_5_digit_division | 0|multiple_choice_grade|0.1800|± |0.0386|
200
+ |bigbench_arithmetic_5_digit_multiplication | 0|multiple_choice_grade|0.2000|± |0.0402|
201
+ |bigbench_arithmetic_5_digit_subtraction | 0|multiple_choice_grade|0.5000|± |0.0503|
202
+ |bigbench_cause_and_effect_one_sentence | 0|multiple_choice_grade|0.5294|± |0.0706|
203
+ |bigbench_cause_and_effect_one_sentence_no_prompt| 0|multiple_choice_grade|0.8627|± |0.0487|
204
+ |bigbench_cause_and_effect_two_sentences | 0|multiple_choice_grade|0.4314|± |0.0700|
205
+ |bigbench_emotions | 0|multiple_choice_grade|0.4750|± |0.0396|
206
+ |bigbench_empirical_judgments | 0|multiple_choice_grade|0.4141|± |0.0498|
207
+ |bigbench_general_knowledge | 0|multiple_choice_grade|0.4429|± |0.0598|
208
+ |bigbench_hhh_alignment_harmless | 0|multiple_choice_grade|0.3793|± |0.0643|
209
+ |bigbench_hhh_alignment_helpful | 0|multiple_choice_grade|0.3220|± |0.0614|
210
+ |bigbench_hhh_alignment_honest | 0|multiple_choice_grade|0.3898|± |0.0640|
211
+ |bigbench_hhh_alignment_other | 0|multiple_choice_grade|0.5581|± |0.0766|
212
+ |bigbench_intent_recognition | 0|multiple_choice_grade|0.2717|± |0.0169|
213
+ |bigbench_misconceptions | 0|multiple_choice_grade|0.5373|± |0.0432|
214
+ |bigbench_paraphrase | 0|multiple_choice_grade|0.5000|± |0.0354|
215
+ |bigbench_sentence_ambiguity | 0|multiple_choice_grade|0.5333|± |0.0649|
216
+ |bigbench_similarities_abstraction | 0|multiple_choice_grade|0.5921|± |0.0567|
217
+
218
+
219
+
220
+ ## Team Members
221
+
222
+ - Aapo Tanskanen, [Hugging Face profile](https://huggingface.co/aapot), [LinkedIn profile](https://www.linkedin.com/in/aapotanskanen/)
223
+ - Rasmus Toivanen, [Hugging Face profile](https://huggingface.co/RASMUS), [LinkedIn profile](https://www.linkedin.com/in/rasmustoivanen/)
224
+
225
+ Feel free to contact us for more details 🤗