Upload 5 files
Browse files- config.json +30 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +7 -0
- tokenizer_config.json +5 -0
- vocab.txt +446 -0
config.json
ADDED
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "/sujin/Models/SaProt/SaProt_1.3B_AF2_start",
|
3 |
+
"architectures": [
|
4 |
+
"EsmForMaskedLM"
|
5 |
+
],
|
6 |
+
"attention_probs_dropout_prob": 0.1,
|
7 |
+
"classifier_dropout": null,
|
8 |
+
"emb_layer_norm_before": false,
|
9 |
+
"esmfold_config": null,
|
10 |
+
"hidden_act": "gelu",
|
11 |
+
"hidden_dropout_prob": 0.1,
|
12 |
+
"hidden_size": 1280,
|
13 |
+
"initializer_range": 0.02,
|
14 |
+
"intermediate_size": 5120,
|
15 |
+
"is_folding_model": false,
|
16 |
+
"layer_norm_eps": 1e-05,
|
17 |
+
"mask_token_id": 4,
|
18 |
+
"max_position_embeddings": 1026,
|
19 |
+
"model_type": "esm",
|
20 |
+
"num_attention_heads": 20,
|
21 |
+
"num_hidden_layers": 66,
|
22 |
+
"pad_token_id": 1,
|
23 |
+
"position_embedding_type": "rotary",
|
24 |
+
"token_dropout": true,
|
25 |
+
"torch_dtype": "float32",
|
26 |
+
"transformers_version": "4.28.0",
|
27 |
+
"use_cache": true,
|
28 |
+
"vocab_list": null,
|
29 |
+
"vocab_size": 446
|
30 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:66c3cc1257f4c048eba48639373b9071f959be3f8417f9b8ba56be986c0563d6
|
3 |
+
size 5204098228
|
special_tokens_map.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"cls_token": "<cls>",
|
3 |
+
"eos_token": "<eos>",
|
4 |
+
"mask_token": "<mask>",
|
5 |
+
"pad_token": "<pad>",
|
6 |
+
"unk_token": "<unk>"
|
7 |
+
}
|
tokenizer_config.json
ADDED
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"clean_up_tokenization_spaces": true,
|
3 |
+
"model_max_length": 1000000000000000019884624838656,
|
4 |
+
"tokenizer_class": "EsmTokenizer"
|
5 |
+
}
|
vocab.txt
ADDED
@@ -0,0 +1,446 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
<cls>
|
2 |
+
<pad>
|
3 |
+
<eos>
|
4 |
+
<unk>
|
5 |
+
<mask>
|
6 |
+
Ap
|
7 |
+
Ay
|
8 |
+
An
|
9 |
+
Aw
|
10 |
+
Ar
|
11 |
+
Aq
|
12 |
+
Ah
|
13 |
+
Ag
|
14 |
+
Ad
|
15 |
+
Al
|
16 |
+
Av
|
17 |
+
At
|
18 |
+
Am
|
19 |
+
Af
|
20 |
+
As
|
21 |
+
Aa
|
22 |
+
Ae
|
23 |
+
Ai
|
24 |
+
Ak
|
25 |
+
Ac
|
26 |
+
A#
|
27 |
+
Cp
|
28 |
+
Cy
|
29 |
+
Cn
|
30 |
+
Cw
|
31 |
+
Cr
|
32 |
+
Cq
|
33 |
+
Ch
|
34 |
+
Cg
|
35 |
+
Cd
|
36 |
+
Cl
|
37 |
+
Cv
|
38 |
+
Ct
|
39 |
+
Cm
|
40 |
+
Cf
|
41 |
+
Cs
|
42 |
+
Ca
|
43 |
+
Ce
|
44 |
+
Ci
|
45 |
+
Ck
|
46 |
+
Cc
|
47 |
+
C#
|
48 |
+
Dp
|
49 |
+
Dy
|
50 |
+
Dn
|
51 |
+
Dw
|
52 |
+
Dr
|
53 |
+
Dq
|
54 |
+
Dh
|
55 |
+
Dg
|
56 |
+
Dd
|
57 |
+
Dl
|
58 |
+
Dv
|
59 |
+
Dt
|
60 |
+
Dm
|
61 |
+
Df
|
62 |
+
Ds
|
63 |
+
Da
|
64 |
+
De
|
65 |
+
Di
|
66 |
+
Dk
|
67 |
+
Dc
|
68 |
+
D#
|
69 |
+
Ep
|
70 |
+
Ey
|
71 |
+
En
|
72 |
+
Ew
|
73 |
+
Er
|
74 |
+
Eq
|
75 |
+
Eh
|
76 |
+
Eg
|
77 |
+
Ed
|
78 |
+
El
|
79 |
+
Ev
|
80 |
+
Et
|
81 |
+
Em
|
82 |
+
Ef
|
83 |
+
Es
|
84 |
+
Ea
|
85 |
+
Ee
|
86 |
+
Ei
|
87 |
+
Ek
|
88 |
+
Ec
|
89 |
+
E#
|
90 |
+
Fp
|
91 |
+
Fy
|
92 |
+
Fn
|
93 |
+
Fw
|
94 |
+
Fr
|
95 |
+
Fq
|
96 |
+
Fh
|
97 |
+
Fg
|
98 |
+
Fd
|
99 |
+
Fl
|
100 |
+
Fv
|
101 |
+
Ft
|
102 |
+
Fm
|
103 |
+
Ff
|
104 |
+
Fs
|
105 |
+
Fa
|
106 |
+
Fe
|
107 |
+
Fi
|
108 |
+
Fk
|
109 |
+
Fc
|
110 |
+
F#
|
111 |
+
Gp
|
112 |
+
Gy
|
113 |
+
Gn
|
114 |
+
Gw
|
115 |
+
Gr
|
116 |
+
Gq
|
117 |
+
Gh
|
118 |
+
Gg
|
119 |
+
Gd
|
120 |
+
Gl
|
121 |
+
Gv
|
122 |
+
Gt
|
123 |
+
Gm
|
124 |
+
Gf
|
125 |
+
Gs
|
126 |
+
Ga
|
127 |
+
Ge
|
128 |
+
Gi
|
129 |
+
Gk
|
130 |
+
Gc
|
131 |
+
G#
|
132 |
+
Hp
|
133 |
+
Hy
|
134 |
+
Hn
|
135 |
+
Hw
|
136 |
+
Hr
|
137 |
+
Hq
|
138 |
+
Hh
|
139 |
+
Hg
|
140 |
+
Hd
|
141 |
+
Hl
|
142 |
+
Hv
|
143 |
+
Ht
|
144 |
+
Hm
|
145 |
+
Hf
|
146 |
+
Hs
|
147 |
+
Ha
|
148 |
+
He
|
149 |
+
Hi
|
150 |
+
Hk
|
151 |
+
Hc
|
152 |
+
H#
|
153 |
+
Ip
|
154 |
+
Iy
|
155 |
+
In
|
156 |
+
Iw
|
157 |
+
Ir
|
158 |
+
Iq
|
159 |
+
Ih
|
160 |
+
Ig
|
161 |
+
Id
|
162 |
+
Il
|
163 |
+
Iv
|
164 |
+
It
|
165 |
+
Im
|
166 |
+
If
|
167 |
+
Is
|
168 |
+
Ia
|
169 |
+
Ie
|
170 |
+
Ii
|
171 |
+
Ik
|
172 |
+
Ic
|
173 |
+
I#
|
174 |
+
Kp
|
175 |
+
Ky
|
176 |
+
Kn
|
177 |
+
Kw
|
178 |
+
Kr
|
179 |
+
Kq
|
180 |
+
Kh
|
181 |
+
Kg
|
182 |
+
Kd
|
183 |
+
Kl
|
184 |
+
Kv
|
185 |
+
Kt
|
186 |
+
Km
|
187 |
+
Kf
|
188 |
+
Ks
|
189 |
+
Ka
|
190 |
+
Ke
|
191 |
+
Ki
|
192 |
+
Kk
|
193 |
+
Kc
|
194 |
+
K#
|
195 |
+
Lp
|
196 |
+
Ly
|
197 |
+
Ln
|
198 |
+
Lw
|
199 |
+
Lr
|
200 |
+
Lq
|
201 |
+
Lh
|
202 |
+
Lg
|
203 |
+
Ld
|
204 |
+
Ll
|
205 |
+
Lv
|
206 |
+
Lt
|
207 |
+
Lm
|
208 |
+
Lf
|
209 |
+
Ls
|
210 |
+
La
|
211 |
+
Le
|
212 |
+
Li
|
213 |
+
Lk
|
214 |
+
Lc
|
215 |
+
L#
|
216 |
+
Mp
|
217 |
+
My
|
218 |
+
Mn
|
219 |
+
Mw
|
220 |
+
Mr
|
221 |
+
Mq
|
222 |
+
Mh
|
223 |
+
Mg
|
224 |
+
Md
|
225 |
+
Ml
|
226 |
+
Mv
|
227 |
+
Mt
|
228 |
+
Mm
|
229 |
+
Mf
|
230 |
+
Ms
|
231 |
+
Ma
|
232 |
+
Me
|
233 |
+
Mi
|
234 |
+
Mk
|
235 |
+
Mc
|
236 |
+
M#
|
237 |
+
Np
|
238 |
+
Ny
|
239 |
+
Nn
|
240 |
+
Nw
|
241 |
+
Nr
|
242 |
+
Nq
|
243 |
+
Nh
|
244 |
+
Ng
|
245 |
+
Nd
|
246 |
+
Nl
|
247 |
+
Nv
|
248 |
+
Nt
|
249 |
+
Nm
|
250 |
+
Nf
|
251 |
+
Ns
|
252 |
+
Na
|
253 |
+
Ne
|
254 |
+
Ni
|
255 |
+
Nk
|
256 |
+
Nc
|
257 |
+
N#
|
258 |
+
Pp
|
259 |
+
Py
|
260 |
+
Pn
|
261 |
+
Pw
|
262 |
+
Pr
|
263 |
+
Pq
|
264 |
+
Ph
|
265 |
+
Pg
|
266 |
+
Pd
|
267 |
+
Pl
|
268 |
+
Pv
|
269 |
+
Pt
|
270 |
+
Pm
|
271 |
+
Pf
|
272 |
+
Ps
|
273 |
+
Pa
|
274 |
+
Pe
|
275 |
+
Pi
|
276 |
+
Pk
|
277 |
+
Pc
|
278 |
+
P#
|
279 |
+
Qp
|
280 |
+
Qy
|
281 |
+
Qn
|
282 |
+
Qw
|
283 |
+
Qr
|
284 |
+
Qq
|
285 |
+
Qh
|
286 |
+
Qg
|
287 |
+
Qd
|
288 |
+
Ql
|
289 |
+
Qv
|
290 |
+
Qt
|
291 |
+
Qm
|
292 |
+
Qf
|
293 |
+
Qs
|
294 |
+
Qa
|
295 |
+
Qe
|
296 |
+
Qi
|
297 |
+
Qk
|
298 |
+
Qc
|
299 |
+
Q#
|
300 |
+
Rp
|
301 |
+
Ry
|
302 |
+
Rn
|
303 |
+
Rw
|
304 |
+
Rr
|
305 |
+
Rq
|
306 |
+
Rh
|
307 |
+
Rg
|
308 |
+
Rd
|
309 |
+
Rl
|
310 |
+
Rv
|
311 |
+
Rt
|
312 |
+
Rm
|
313 |
+
Rf
|
314 |
+
Rs
|
315 |
+
Ra
|
316 |
+
Re
|
317 |
+
Ri
|
318 |
+
Rk
|
319 |
+
Rc
|
320 |
+
R#
|
321 |
+
Sp
|
322 |
+
Sy
|
323 |
+
Sn
|
324 |
+
Sw
|
325 |
+
Sr
|
326 |
+
Sq
|
327 |
+
Sh
|
328 |
+
Sg
|
329 |
+
Sd
|
330 |
+
Sl
|
331 |
+
Sv
|
332 |
+
St
|
333 |
+
Sm
|
334 |
+
Sf
|
335 |
+
Ss
|
336 |
+
Sa
|
337 |
+
Se
|
338 |
+
Si
|
339 |
+
Sk
|
340 |
+
Sc
|
341 |
+
S#
|
342 |
+
Tp
|
343 |
+
Ty
|
344 |
+
Tn
|
345 |
+
Tw
|
346 |
+
Tr
|
347 |
+
Tq
|
348 |
+
Th
|
349 |
+
Tg
|
350 |
+
Td
|
351 |
+
Tl
|
352 |
+
Tv
|
353 |
+
Tt
|
354 |
+
Tm
|
355 |
+
Tf
|
356 |
+
Ts
|
357 |
+
Ta
|
358 |
+
Te
|
359 |
+
Ti
|
360 |
+
Tk
|
361 |
+
Tc
|
362 |
+
T#
|
363 |
+
Vp
|
364 |
+
Vy
|
365 |
+
Vn
|
366 |
+
Vw
|
367 |
+
Vr
|
368 |
+
Vq
|
369 |
+
Vh
|
370 |
+
Vg
|
371 |
+
Vd
|
372 |
+
Vl
|
373 |
+
Vv
|
374 |
+
Vt
|
375 |
+
Vm
|
376 |
+
Vf
|
377 |
+
Vs
|
378 |
+
Va
|
379 |
+
Ve
|
380 |
+
Vi
|
381 |
+
Vk
|
382 |
+
Vc
|
383 |
+
V#
|
384 |
+
Wp
|
385 |
+
Wy
|
386 |
+
Wn
|
387 |
+
Ww
|
388 |
+
Wr
|
389 |
+
Wq
|
390 |
+
Wh
|
391 |
+
Wg
|
392 |
+
Wd
|
393 |
+
Wl
|
394 |
+
Wv
|
395 |
+
Wt
|
396 |
+
Wm
|
397 |
+
Wf
|
398 |
+
Ws
|
399 |
+
Wa
|
400 |
+
We
|
401 |
+
Wi
|
402 |
+
Wk
|
403 |
+
Wc
|
404 |
+
W#
|
405 |
+
Yp
|
406 |
+
Yy
|
407 |
+
Yn
|
408 |
+
Yw
|
409 |
+
Yr
|
410 |
+
Yq
|
411 |
+
Yh
|
412 |
+
Yg
|
413 |
+
Yd
|
414 |
+
Yl
|
415 |
+
Yv
|
416 |
+
Yt
|
417 |
+
Ym
|
418 |
+
Yf
|
419 |
+
Ys
|
420 |
+
Ya
|
421 |
+
Ye
|
422 |
+
Yi
|
423 |
+
Yk
|
424 |
+
Yc
|
425 |
+
Y#
|
426 |
+
#p
|
427 |
+
#y
|
428 |
+
#n
|
429 |
+
#w
|
430 |
+
#r
|
431 |
+
#q
|
432 |
+
#h
|
433 |
+
#g
|
434 |
+
#d
|
435 |
+
#l
|
436 |
+
#v
|
437 |
+
#t
|
438 |
+
#m
|
439 |
+
#f
|
440 |
+
#s
|
441 |
+
#a
|
442 |
+
#e
|
443 |
+
#i
|
444 |
+
#k
|
445 |
+
#c
|
446 |
+
##
|