Hugging Face's logo

m-a-p
/
ChatMusician

Text Generation
Transformers
PyTorch
Safetensors
English
llama
text-generation-inference
Inference Endpoints
Model card Files Community
5
agent404 commited on
Commit
a8b376d
•
1 Parent(s): f958638

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +358 -1
README.md CHANGED
@@ -1,3 +1,360 @@
1
  ---
2
- license: cc
 
 
 
 
 
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
+ license: mit
3
+ language:
4
+ - en
5
+ metrics:
6
+ - accuracy
7
+ pipeline_tag: text-generation
8
  ---
9
+ # 🎼 ChatMusician: Fostering Intrinsic Musical Abilities Into LLM
10
+
11
+ [**🌐 DemoPage**](https://ezmonyi.github.io/ChatMusician/) | [**🤗 Dataset**](https://huggingface.co/datasets/m-a-p/MusicPile) | [**🤗 Benchmark**](https://huggingface.co/datasets/m-a-p/MusicTheoryBench) | [**📖 arXiv**](http://arxiv.org/abs/2402.16153) | [**Code**](https://github.com/hf-lin/ChatMusician)
12
+
13
+ ## 🔔News
14
+ - **🔥[2023-12-10]: The release of ChatMusician's demo, code, model, data, and benchmark. 😆**
15
+ - [2023-11-30]: Checkout another awesome project [MMMU](https://huggingface.co/datasets/MMMU/MMMU/) that includes multimodal music reasoning.
16
+
17
+ ## Introduction
18
+
19
+ While Large Language Models (LLMs) demonstrate impressive capabilities in text generation,
20
+ we find that their ability has yet to be generalized to music, humanity’s creative language.
21
+ We introduce **ChatMusician**, **an open-source LLM that integrates intrinsic musical abilities**.
22
+
23
+ It is based on continual pre-training and finetuning LLaMA2 on a text-compatible music representation, ABC notation, and the music is treated as a second language. ChatMusician can understand and generate music with a pure text tokenizer without any external multi-modal neural structures or tokenizers. Interestingly, endowing musical abilities does not harm language abilities, even achieving a slightly higher MMLU score. Our model is capable of composing well-structured, full-length music, conditioned on texts, chords, melodies, motifs, musical forms, etc, surpassing GPT-4 baseline. On our meticulously curated college-level music understanding benchmark, MusicTheoryBench, ChatMusician surpasses LLaMA2 and GPT-3.5 on zero-shot setting by a noticeable
24
+ margin. Our work reveals that LLMs can be an excellent compressor for music, but there remains significant territory to be conquered. Code, data, model, and benchmark are open-sourced.
25
+
26
+ <!-- <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/5fd6f670053c8345eddc1b68/8NSONUjIF7KGUCfwzPCd9.mpga"></audio> -->
27
+
28
+ ## Training Data
29
+
30
+ ChatMusician is pretrained on the 🤗 [MusicPile](https://huggingface.co/datasets/m-a-p/MusicPile), which is the first pretraining corpus for **developing musical abilities** in large language models. Check out the dataset card for more details.
31
+ And supervised finetuned on 1.1M samples(2:1 ratio between music scores
32
+ and music knowledge & music summary data) from MusicPile. Check our [paper](http://arxiv.org/abs/2402.16153) for more details.
33
+
34
+ ## Training Procedure
35
+
36
+ We initialized a fp16-precision ChatMusician-Base from the LLaMA2-7B-Base weights, and applied a continual pre-training plus fine-tuning pipeline. LoRA adapters were integrated into the attention and MLP layers, with additional training on embeddings and all linear layers. The maximum sequence length
37
+ was 2048. We utilized 16 80GB-A800 GPUs for one epoch pre-training and 8 32GB-V100 GPUs for two epoch fine-tuning. DeepSpeed was employed for memory efficiency, and the AdamW optimizer was used with a 1e-4 learning rate and a 5% warmup cosine scheduler. Gradient clipping was set at 1.0. The LoRA parameters dimension, alpha, and
38
+ dropout were set to 64, 16, and 0.1, with a batch size of 8.
39
+
40
+ ## Evaluation
41
+
42
+ 1. Music understanding abilities are evaluated on the [MusicTheoryBench](https://huggingface.co/datasets/m-a-p/MusicTheoryBench).
43
+ 2. General language abilities of ChatMusician are evaluated on the [Massive Multitask Language Understanding (MMLU) dataset](https://huggingface.co/datasets/lukaemon/mmlu).
44
+
45
+
46
+ ## Usage
47
+
48
+ You can use the models through Huggingface's Transformers library. Check our Github repo for more advanced use: [https://github.com/hf-lin/ChatMusician](https://github.com/hf-lin/ChatMusician)
49
+
50
+
51
+ ## Prompt Format
52
+
53
+ **Our model produces symbolic music(ABC notation) well in the following prompts.** Here are some musical tasks.
54
+
55
+ ### Function: Text2music
56
+
57
+ ```
58
+ Prompt:
59
+ Using ABC notation, recreate the given text as a musical score.
60
+ Meter C
61
+ Notes The parts are commonly interchanged.
62
+ Transcription 1997 by John Chambers
63
+ Key D
64
+ Note Length 1/8
65
+ Rhythm reel
66
+
67
+ Response:
68
+ X:1
69
+ L:1/8
70
+ M:2/2
71
+ K:D
72
+ "A7" FE |"D" D2 FA dAFD |"A7" CDEF G2 FE |"D" D2 FA dcdf |"E7" edcB"A7" AGFE | |"D" D2 FA dAFD |
73
+ "A7" CDEF G2 FG |"D" AdcB"A7" AGFE |"D" D2 F2 D2 :: de |"D" fefg fedc |"Em" B^ABc B2 cd |
74
+ "E7" e^def e=dcB |"A" A2 ce aABc | |"D" dcde fdAF |"G" GFGA BdcB |"D" AdcB"A7" Agec |
75
+ "D" d2 f2 d2 :|
76
+ ```
77
+ ```
78
+ Prompt:
79
+ Develop a tune influenced by Bach's compositions.
80
+
81
+ Response:
82
+ X:1
83
+ X:1
84
+ T:Music21 Fragment
85
+ T:Music21 Fragment
86
+ T:Music21
87
+ C:Music21
88
+ %%score 1 2 3 4
89
+ L:1/8
90
+ M:4/4
91
+ K:C
92
+ V:1 treble nm="Piano" snm="Pno"
93
+ %%MIDI program 0
94
+ %%MIDI control 7 100
95
+ %%MIDI control 10 64
96
+ L:1/4
97
+ V:2 treble nm="Piano" snm="Pno"
98
+ %%MIDI channel 3
99
+ %%MIDI program 0
100
+ %%MIDI control 7 100
101
+ %%MIDI control 10 64
102
+ V:3 bass nm="Piano" snm="Pno"
103
+ %%MIDI channel 4
104
+ %%MIDI program 0
105
+ %%MIDI control 7 100
106
+ %%MIDI control 10 64
107
+ V:4 bass nm="Piano" snm="Pno"
108
+ %%MIDI channel 5
109
+ %%MIDI program 0
110
+ %%MIDI control 7 100
111
+ %%MIDI control 10 64
112
+ V:1
113
+ z3 c | B A G F | E D !fermata!E c | d e d c | !fermata!B2 z c | B A G F | E D !fermata!E c | %7
114
+ d e d c | !fermata!B2 z e | f e d c | B A !fermata!B B | c d e d | !fermata!c2 z e | d c B A | %14
115
+ G A B c | d e d c | !fermata!B2 z e | d c B A | G A B c | d e d c | !fermata!c2 z B | A G F E | %22
116
+ D C !fermata!D c | B c d e | d c B A | !fermata!G2 z c | B A G F | E D !fermata!E c | d e d c | %29
117
+ !fermata!B2 z e | d c B A | G A B c | d e d c | !fermata!c2 z e | d c B A | G A B c | d e d c | %36
118
+ !fermata!c2 z B | A G F E | D C !fermata!D c | B c d e | d c B A | !fermata!G2 z2 |] %42
119
+ V:2
120
+ z6 CD | EF ED C2 CD | EF ED !fermata!C2 CD | EF GF ED C2 | !fermata!C6 CD | EF ED C2 CD | %6
121
+ EF ED !fermata!C2 CD | EF GF ED C2 | !fermata!C6 CD | EF GA GF ED | C2 DE !fermata!F2 DC | %11
122
+ B,C DB, C2 CD | !fermata!E6 CD | EF GA GF ED | C2 DE !fermata!F2 DC | B,C DB, C2 CD | %16
123
+ !fermata!E6 CD | EF GA GF ED | C2 DE !fermata!F2 DC | B,C DB, C2 CD | !fermata!E6 B,C | %21
124
+ DE ^FE D=C B,A, | G,^F, G,2 !fermata!G,2 =F,2 | G,2 A,B, CD EF | !fermata!G6 B,C | DE ^FE D=C B,A, | %26
125
+ G,^F, G,2 !fermata!G,2 =F,2 | G,2 A,B, CD EF | !fermata!G6 B,C | DE ^FE D=C B,A, | %30
126
+ G,^F, G,2 !fermata!G,2 =F,2 | G,2 A,B, CD EF | !fermata!G6 B,C | DE ^FE D=C B,A, | %34
127
+ G,^F, G,2 !fermata!G,2 =F,2 | G,2 A,B, CD EF | !fermata!G6 z2 |] %37
128
+ V:3
129
+ z6 G,2 | A,2 G,2 G,2 G,2 | G,2 G,2 !fermata!G,2 G,2 | G,2 A,2 B,2 C2 | !fermata!C6 G,2 | %6
130
+ A,2 G,2 G,2 G,2 | G,2 G,2 !fermata!G,2 G,2 | G,2 A,2 B,2 C2 | !fermata!C6 G,2 | G,2 A,2 B,2 C2 | %11
131
+ C2 B,2 !fermata!C2 B,2 | C2 D2 E2 D2 | !fermata!C6 G,2 | A,2 G,2 G,2 G,2 | G,2 G,2 !fermata!G,2 G,2 | %16
132
+ G,2 A,2 B,2 C2 | !fermata!C6 G,2 | G,2 A,2 B,2 C2 | C2 B,2 !fermata!C2 B,2 | C2 D2 E2 D2 | %21
133
+ !fermata!C6 B,2 | A,2 G,2 G,2 G,2 | G,2 G,2 !fermata!G,2 G,2 | G,2 A,2 B,2 C2 | !fermata!C6 B,2 | %26
134
+ A,2 G,2 G,2 G,2 | G,2 G,2 !fermata!G,2 G,2 | G,2 A,2 B,2 C2 | !fermata!C6 B,2 | A,2 G,2 G,2 G,2 | %31
135
+ G,2 G,2 !fermata!G,2 G,2 | G,2 A,2 B,2 C2 | !fermata!C6 z2 |] %34
136
+ V:4
137
+ z6"C" C,2 |"Am" A,,2"Em" E,2"C" C,2"C" C,2 |"Am" A,,2"Em" E,2"C" !fermata!C,2"C" C,2 | %3
138
+ "G" G,,2"Am" A,,2"G" B,,2"C" C,2 |"C" !fermata!C,6"C" C,2 |"Am" A,,2"Em" E,2"C" C,2"C" C,2 | %6
139
+ "Am" A,,2"Em" E,2"C" !fermata!C,2"C" C,2 |"G" G,,2"Am" A,,2"G" B,,2"C" C,2 |"C" !fermata!C,6"C" C,2 | %9
140
+ "G" G,,2"Am" A,,2"G" B,,2"C" C,2 |"C" C,2"G" G,,2"C" !fermata!C,2"C" C,2 | %11
141
+ "C" C,2"G" G,,2"Am" A,,2"G" B,,2 |"C" !fermata!C,6"C" C,2 |"G" G,,2"Am" A,,2"G" B,,2"C" C,2 | %14
142
+ "C" C,2"G" G,,2"C" !fermata!C,2"C" C,2 |"G" G,,2"Am" A,,2"G" B,,2"C" C,2 |"C" !fermata!C,6"C" C,2 | %17
143
+ "G" G,,2"Am" A,,2"G" B,,2"C" C,2 |"C" C,2"G" G,,2"C" !fermata!C,2"C" C,2 | %19
144
+ "G" G,,2"Am" A,,2"G" B,,2"C" C,2 |"C" !fermata!C,6"E" E,2 |"Am" A,,2"Em" E,2"C" C,2"C" C,2 | %
145
+ ```
146
+
147
+ ### Function: Melody Harmonization
148
+
149
+ ```
150
+ Prompt:
151
+ Construct smooth-flowing chord progressions for the supplied music.
152
+
153
+ |: BA | G2 g2"^(C)" edeg | B2 BA"^(D7)" BcBA | G2 g2 edeg | dBAG A2 BA |
154
+ G2 g2"^(C)" edeg | B2 BA B2 d2 | e2 ef e2 (3def | gedB A2 :: BA | G2 BG dGBe |
155
+ dBBA"^(D7)" B3 A | G2 BG dGBe | dBAG A4 | G2 BG dGBe | dBBA B3 d |
156
+ e2 ef e2 (3def | gedB A2 :|
157
+
158
+ Response:
159
+ X:1
160
+ L:1/8
161
+ M:2/2
162
+ K:G
163
+ |: BA |"G" G2 g2"^(C)" edeg |"G" B2 BA"^(D7)" BcBA |"G" G2 g2"Em" edeg |"Am" A2 AG"D7" AcBA |
164
+ "G" G2 g2"^(C)" edeg |"G" B2 BA B2 d2 |"C" e2 ef e2 (3def |"D7" gedB A2 :: BA |"G" G2 BG dGBG |
165
+ "G" B2 BA"^(D7)" BcBA |"G" G2 BG"Em" dGBG |"Am" A2 AG"D7" AcBA |"G" G2 BG dGBe | dBBA B2 d2 |
166
+ "C" e2 ef e2 (3def |"D7" gedB A2 :|
167
+ ```
168
+ ```
169
+ Prompt:
170
+ Develop a series of chord pairings that amplify the harmonious elements in the given music piece.
171
+
172
+ E |: EAA ABc | Bee e2 d | cBA ABc | BEE E2 D | EAA ABc | Bee e2 d |
173
+ cBA ^GAB |1 A2 A A2 E :|2 A2 A GAB || c3 cdc | Bgg g2 ^g | aed cBA |
174
+ ^GAB E^F^G | A^GA BAB | cde fed | cBA ^GAB |1 A2 A GAB :|2 \n A3 A2 ||
175
+
176
+ Response:
177
+ X:1
178
+ L:1/8
179
+ M:6/8
180
+ K:Amin
181
+ E |:"Am" EAA ABc |"E7" Bee e2 d |"Am" cBA ABc |"E7" BGE E2 E |
182
+ "Am" EAA ABc |"E7" Bee e2 d |"Am" cBA"E7" ^GAB |1"Am" A2 A A2 E :|2
183
+ "Am" A2 A GAB ||"C" c3 cdc |"G" Bgg g2 ^g |"Am" aed cBA |
184
+ "E7" ^GAB E^F^G |"Am" A^GA"E7" BAB |"Am" cde"Dm" fed |
185
+ "Am" cBA"E7" ^GAB |1"Am" A2 A GAB :|2 "Am" A2 A A2 ||
186
+ ```
187
+
188
+ ### Function: Chord Conditioned Music Generation
189
+ ```
190
+ Prompt:
191
+ Develop a musical piece using the given chord progression. 'Dm', 'C', 'Dm', 'Dm', 'C', 'Dm', 'C', 'Dm'
192
+
193
+ Response:
194
+ X:1
195
+ L:1/8
196
+ M:6/8
197
+ K:F
198
+ z4 z D |"Dm" A2 A"C" G2 F |"Dm" FE D D2 D |
199
+ "Dm" A2 A"C" G2 A |"Dm" D3 D2 D |"Dm" A2 A"C" G2 F |
200
+ "Dm" FE D D2 D |"Dm" F2 G"C" E2 C |"Dm" D3 D2 D |
201
+ "F" F2 F F2 c |"Bb" d2 c"^(Gm)" B2 A |
202
+ "Gm" G2 G G2 B |"Am" A2 G"A7" FE D |"Dm" F2 F F2 c |
203
+ "Bb" d2 c"^(Gm)" B2 A |"Gm" G2 G"A7" A2 G |
204
+ "Dm" FE D D2 z |]
205
+ ```
206
+
207
+ ### Function: Musical Form Conditioned Music Generation
208
+
209
+ ```
210
+ Prompt:
211
+ Develop a composition by incorporating elements from the given melodic structure.
212
+
213
+ Ternary, Sectional: Verse/Chorus/Bridge
214
+
215
+ Response:
216
+ X:1
217
+ L:1/8
218
+ M:6/8
219
+ K:Gmin
220
+ |:"Gm" GAB AG^F | GAB c2 c | d2 d cdB | A2 d cBA |
221
+ GAB AG^F | GAB c2 c | d2 d cdB |1 AG^F G2 D :|2
222
+ AG^F GAB |: c2 c cBA | c2 c cBA | c2 c cBA |
223
+ ABA GAB | c2 c cBA | c2 c cBA | d2 d cdB |1
224
+ AG^F GAB :|2 AG^F G2 A |:"Bbm" B3 dcB | f2 e dcB |
225
+ c2 f c2 c | cfc cBA | ABB dcB | f2 e dcB |
226
+ d2 d edc |1 cBA B2 B :|2 cBA G3 ||
227
+ ```
228
+
229
+ ### Function: Motif and Form Conditioned Music Generation
230
+
231
+ ```
232
+ Text Input:
233
+ Create music by incorporating the assigned motif into the predetermined musical arrangement.
234
+
235
+ Musical Form Input: Only One Section
236
+
237
+ ABC Notation Music Input:
238
+ X:1
239
+ L:1/8
240
+ M:9/8
241
+ K:Emin
242
+ vB2 E E2 F G2 A
243
+
244
+ Response:
245
+ X:1
246
+ L:1/8
247
+ M:9/8
248
+ K:Emin
249
+ |:"Em" vB2 E E2 F G2 A |"D" B2 e e2 ^d e3 |
250
+ "Bm" B2 E E2 F G2 A |"G" B2 G A2 F"C" G2 E |
251
+ "Em" vB2 E E2 F G2 A |"D" B2 e e2 ^d"C" e3 |
252
+ "Em" ve2 B"^Av" A2 F"Bm" G2 A |1"G" B2 G"D" A2 F"Em" G2 E :|
253
+ 2"G" B2 G"D" A2 F"G" G3 |"^(C)" ABc"^(D)" ded"G" dBG |
254
+ "^(C)" cde"D" dcB"G" A2 G |"^(C)" E2 E"Am" ABc"D" def |
255
+ "Em" g2 G"D" FAF"G" G3 |"^(C)" ABc"^(D)" ded"G" dBG |
256
+ "^(C)" cde"D" dcB"G" A2 G |"Em" e2 B"^Av" A2 F"Bm" G2 A |
257
+ "G" B2 G"D" A2 F"Em" G2 E |]
258
+ ```
259
+
260
+ ### Function: Music Understanding
261
+
262
+ ```
263
+ Prompt:
264
+ Investigate the aspects of this musical work and convey its structural organization using suitable musical words.
265
+
266
+ X:1
267
+ L:1/8
268
+ M:2/2
269
+ K:G
270
+ G2 dG BGdG | G2 dc BAGB | A2 eA cAeA | A2 ed cAFA |
271
+ G2 dG BGdG | G2 dc BAGB | ABcd efge |1 aged cAFA :|2
272
+ aged ^cdef |: g3 f g2 ef | gedc BA G2 | eaag agea |
273
+ aged ^cdef | g3 f g2 ef |gedc BAGB | ABcd efge |1
274
+ aged ^cdef :|2 aged cAFA |:"^variations:" G2 BG dGBA |
275
+ G2 dG BAGB | A2 cA eAcA | A2 ed cAFA | G2 BG dGBA |
276
+ G2 dc BAGB | ABcd efge |1 aged cAFA :|2 aged ^cdef |:
277
+ g2 af g2 ef | gedc BAGB | Aaag ageg | aged ^cdef |
278
+ gbaf g2 ef | gedc BAGB | ABcd efge |1
279
+ aged ^cdef :|2 aged cAFA ||
280
+
281
+ Response:
282
+ Compound Binary
283
+ ```
284
+
285
+ ```
286
+ Prompt:
287
+ Analyze the musical work and pinpoint the consistent melodic element in every section.
288
+
289
+ X:1
290
+ L:1/8
291
+ M:4/4
292
+ K:G
293
+ ge | d2 G2 cBAG | d2 G2 cBAG | e2 A2 ABcd | edcB A2 Bc |
294
+ d2 cB g2 fe | edcB cBAG | BAGE DEGA | B2 G2 G2 :: ga |
295
+ b2 gb a2 fa | g2 eg edcB | e2 A2 ABcd | edcB A2 ga |
296
+ b2 gb a2 fa | g2 eg edcB | cBAG DEGA | B2 G2 G2 :|
297
+
298
+ Response:
299
+ X:1
300
+ L:1/8
301
+ M:4/4
302
+ K:G
303
+ ge d2 G2 cBAG d2 G2 cBAG
304
+ ```
305
+
306
+ ## CLI demo
307
+ ```
308
+ from transformers import AutoTokenizer, AutoModelForCausalLM, GenerationConfig
309
+ from string import Template
310
+ prompt_template = Template("Human: ${inst} </s> Assistant: ")
311
+
312
+ tokenizer = AutoTokenizer.from_pretrained("m-a-p/ChatMusician-v1-sft-78k", trust_remote_code=True)
313
+ model = AutoModelForCausalLM.from_pretrained("m-a-p/ChatMusician-v1-sft-78k", trust_remote_code=True).eval()
314
+ model.cuda()
315
+ generation_config = GenerationConfig(
316
+ temperature=0.2,
317
+ top_k=40,
318
+ top_p=0.9,
319
+ do_sample=True,
320
+ num_beams=1,
321
+ repetition_penalty=1.1,
322
+ min_new_tokens=10,
323
+ max_new_tokens=1536
324
+ )
325
+
326
+ instruction = """Using ABC notation, recreate the given text as a musical score.
327
+ Meter C
328
+ Notes The parts are commonly interchanged.
329
+ Transcription 1997 by John Chambers
330
+ Key D
331
+ Note Length 1/8
332
+ Rhythm reel
333
+ """
334
+
335
+ prompt = prompt_template.safe_substitute({"inst": instruction})
336
+ inputs = tokenizer(prompt, return_tensors="pt", add_special_tokens=False)
337
+ response = model.generate(
338
+ input_ids=inputs["input_ids"].to(model.device),
339
+ attention_mask=inputs['attention_mask'].to(model.device),
340
+ eos_token_id=tokenizer.eos_token_id,
341
+ generation_config=generation_config,
342
+ )
343
+ response = tokenizer.decode(response[0][inputs["input_ids"].shape[1]:], skip_special_tokens=True)
344
+ print(response)
345
+ ```
346
+
347
+ ## Intended Uses
348
+ These models are trained for research purposes. They are designed to solve general math problems. They can be used in educational software, tutoring systems, or any application where a solution to a math problem is needed. The models can generate both a chain of thought (CoT) rationale and a program of thought (PoT) rationale, providing a comprehensive solution to a given math problem.
349
+
350
+
351
+ ## Limitations
352
+ We've tried our best to build math generalist models. However, we acknowledge that the models' performance may vary based on the complexity and specifics of the math problem. Still not all mathematical fields can be covered comprehensively.
353
+
354
+
355
+ ## Citation
356
+ If you use the models, data, or code from this project, please cite the original paper:
357
+
358
+ ```
359
+ coming soon.
360
+ ```