PeytonT commited on
Commit
fe55445
·
verified ·
1 Parent(s): 5ea9b02

Update AgentKernel Lite browser BitNet to v11 recommendation answer model

Browse files

Exports recommendation_answer_v11_exact_twopass_from_v9_train_00200_lr2e6 as a model-stack browser BitNet bundle with WASM/WebGPU runtime files.

This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. README.md +3 -8
  2. agentkernel_lite_browser_bitnet_export.json +19 -10
  3. dense/dec_embed_weight.f32.bin +2 -2
  4. dense/dec_norm_bias.f32.bin +1 -1
  5. dense/dec_norm_weight.f32.bin +1 -1
  6. dense/decoder_0_cross_block_n1_weight.f32.bin +1 -1
  7. dense/decoder_0_cross_block_n2_weight.f32.bin +1 -1
  8. dense/decoder_0_self_attn_block_attn__rope_sin.f32.bin +0 -3
  9. dense/decoder_0_self_attn_block_n1_weight.f32.bin +1 -1
  10. dense/decoder_0_self_attn_block_n2_weight.f32.bin +1 -1
  11. dense/decoder_1_cross_block_n1_weight.f32.bin +1 -1
  12. dense/decoder_1_cross_block_n2_weight.f32.bin +1 -1
  13. dense/decoder_1_self_attn_block_attn__rope_cos.f32.bin +0 -3
  14. dense/decoder_1_self_attn_block_attn__rope_sin.f32.bin +0 -3
  15. dense/decoder_1_self_attn_block_n1_weight.f32.bin +1 -1
  16. dense/decoder_1_self_attn_block_n2_weight.f32.bin +1 -1
  17. dense/decoder_2_cross_block_n1_weight.f32.bin +1 -1
  18. dense/decoder_2_cross_block_n2_weight.f32.bin +1 -1
  19. dense/decoder_2_self_attn_block_attn__rope_cos.f32.bin +0 -3
  20. dense/decoder_2_self_attn_block_attn__rope_sin.f32.bin +0 -3
  21. dense/decoder_2_self_attn_block_n1_weight.f32.bin +1 -1
  22. dense/decoder_2_self_attn_block_n2_weight.f32.bin +1 -1
  23. dense/decoder_3_cross_block_n1_weight.f32.bin +1 -1
  24. dense/decoder_3_cross_block_n2_weight.f32.bin +1 -1
  25. dense/decoder_3_self_attn_block_attn__rope_cos.f32.bin +0 -3
  26. dense/decoder_3_self_attn_block_attn__rope_sin.f32.bin +0 -3
  27. dense/decoder_3_self_attn_block_n1_weight.f32.bin +1 -1
  28. dense/decoder_3_self_attn_block_n2_weight.f32.bin +1 -1
  29. dense/decoder_4_cross_block_n1_weight.f32.bin +1 -1
  30. dense/decoder_4_cross_block_n2_weight.f32.bin +1 -1
  31. dense/decoder_4_self_attn_block_attn__rope_cos.f32.bin +0 -3
  32. dense/decoder_4_self_attn_block_attn__rope_sin.f32.bin +0 -3
  33. dense/decoder_4_self_attn_block_n1_weight.f32.bin +1 -1
  34. dense/decoder_4_self_attn_block_n2_weight.f32.bin +1 -1
  35. dense/decoder_5_cross_block_n1_weight.f32.bin +1 -1
  36. dense/decoder_5_cross_block_n2_weight.f32.bin +1 -1
  37. dense/decoder_5_self_attn_block_attn__rope_cos.f32.bin +0 -3
  38. dense/decoder_5_self_attn_block_attn__rope_sin.f32.bin +0 -3
  39. dense/decoder_5_self_attn_block_n1_weight.f32.bin +1 -1
  40. dense/decoder_5_self_attn_block_n2_weight.f32.bin +1 -1
  41. dense/enc_embed_weight.f32.bin +2 -2
  42. dense/enc_norm_bias.f32.bin +1 -1
  43. dense/enc_norm_weight.f32.bin +1 -1
  44. dense/{decoder_0_self_attn_block_attn__rope_cos.f32.bin → enc_pos_embed_weight.f32.bin} +2 -2
  45. dense/encoder_0_n1_weight.f32.bin +1 -1
  46. dense/encoder_0_n2_weight.f32.bin +1 -1
  47. dense/encoder_1_n1_weight.f32.bin +1 -1
  48. dense/encoder_1_n2_weight.f32.bin +1 -1
  49. dense/encoder_2_n1_weight.f32.bin +1 -1
  50. dense/encoder_2_n2_weight.f32.bin +1 -1
README.md CHANGED
@@ -7,18 +7,13 @@ tags:
7
  - encoder-decoder
8
  ---
9
 
10
- https://github.com/peytontolbert/agent_kernel
11
-
12
-
13
- http://peytontolbert.com/agent_kernel/
14
-
15
  # AgentKernel Lite Encoder-Decoder Browser BitNet
16
 
17
  Self-contained browser BitNet export for the AgentKernel Lite chat model.
18
 
19
- - Source bundle: `/data/agentkernel/artifacts/agentkernel_lite_encdec/decoder_repair_respond_clean_from_freezeemb_retrieval2000_train_01500`
20
  - Parameters before BitNet packing: `113507328`
21
- - Final eval loss: `0.9985759307940801`
22
  - Browser entrypoint: `manifest.json`
23
  - Runtime: Model Stack browser BitNet WebGPU encoder-decoder with packed BitNet WASM fallback
24
  - Tokenizer: AgentKernel byte-level BPE attached under `tokenizer/`
@@ -26,7 +21,7 @@ Self-contained browser BitNet export for the AgentKernel Lite chat model.
26
  Web app route after uploading this directory to Hugging Face:
27
 
28
  ```text
29
- ?modelStackManifest=https://huggingface.co/PeytonT/agentkernel-lite-100m-bitnet/resolve/main/manifest.json
30
  ```
31
 
32
  Serving notes: WebGPU is used when available; Safari or other no-WebGPU browsers use the packed BitNet WASM fallback. Large model files are fetched by the browser and cached by the app.
 
7
  - encoder-decoder
8
  ---
9
 
 
 
 
 
 
10
  # AgentKernel Lite Encoder-Decoder Browser BitNet
11
 
12
  Self-contained browser BitNet export for the AgentKernel Lite chat model.
13
 
14
+ - Source bundle: `/data/agentkernel/artifacts/agentkernel_lite_encdec/recommendation_answer_v11_exact_twopass_from_v9_train_00200_lr2e6`
15
  - Parameters before BitNet packing: `113507328`
16
+ - Final eval loss: `0.0016785785555839539`
17
  - Browser entrypoint: `manifest.json`
18
  - Runtime: Model Stack browser BitNet WebGPU encoder-decoder with packed BitNet WASM fallback
19
  - Tokenizer: AgentKernel byte-level BPE attached under `tokenizer/`
 
21
  Web app route after uploading this directory to Hugging Face:
22
 
23
  ```text
24
+ ?modelStackManifest=https://huggingface.co/<org>/<repo>/resolve/main/manifest.json
25
  ```
26
 
27
  Serving notes: WebGPU is used when available; Safari or other no-WebGPU browsers use the packed BitNet WASM fallback. Large model files are fetched by the browser and cached by the app.
agentkernel_lite_browser_bitnet_export.json CHANGED
@@ -1,15 +1,18 @@
1
  {
2
  "artifact_kind": "agentkernel_lite_browser_bitnet_export",
 
 
3
  "dense_tensor_count": 43,
4
- "device": "cuda:1",
5
  "format": "model-stack-browser-bitnet",
6
  "layer_count": 109,
7
- "manifest_path": "/data/agentkernel/artifacts/agentkernel_lite_encdec/decoder_repair_respond_clean_from_freezeemb_retrieval2000_train_01500/browser_bitnet/manifest.json",
8
  "max_seq_len": 1024,
9
  "model": {
10
  "activation": "silu",
 
11
  "attention_bias": false,
12
- "attn_dropout": 0.0,
13
  "attn_impl": "eager",
14
  "d_ff": 2048,
15
  "d_model": 640,
@@ -19,13 +22,16 @@
19
  "kv_cache_paged": true,
20
  "masking": "build_causal_mask",
21
  "max_position_embeddings": 4096,
 
22
  "n_heads": 10,
23
  "n_kv_heads": null,
24
  "n_layers": 6,
25
  "norm": "layer",
26
  "pad_token_id": 0,
27
  "positional": "apply_rotary",
 
28
  "residual": "prenorm",
 
29
  "rms_norm_eps": 1e-06,
30
  "rope_attention_scaling": null,
31
  "rope_scaling_factor": null,
@@ -39,7 +45,9 @@
39
  "version": 1,
40
  "vocab_size": 8207
41
  },
42
- "output_dir": "/data/agentkernel/artifacts/agentkernel_lite_encdec/decoder_repair_respond_clean_from_freezeemb_retrieval2000_train_01500/browser_bitnet",
 
 
43
  "quantization": {
44
  "activation_quant": "none",
45
  "modules": [
@@ -158,10 +166,11 @@
158
  "spin": false,
159
  "weight_opt": "none"
160
  },
161
- "size_bytes": 80235516,
162
- "source_bundle_manifest_path": "/data/agentkernel/artifacts/agentkernel_lite_encdec/decoder_repair_respond_clean_from_freezeemb_retrieval2000_train_01500/agentkernel_lite_encdec_manifest.json",
163
- "source_model_dir": "/data/agentkernel/artifacts/agentkernel_lite_encdec/decoder_repair_respond_clean_from_freezeemb_retrieval2000_train_01500/model",
164
- "source_tokenizer_dir": "/data/agentkernel/artifacts/agentkernel_lite_encdec/decoder_repair_respond_clean_from_freezeemb_retrieval2000_train_01500/tokenizer",
 
165
  "tokenizer": {
166
  "bos_token_id": 1,
167
  "config_path": "tokenizer/tokenizer_config.json",
@@ -173,6 +182,6 @@
173
  "unk_token_id": 3,
174
  "vocab_size": 8207
175
  },
176
- "webapp_manifest_path": "/data/repository_library/exports/agent_kernel/models/agentkernel_lite_100m_bitnet_15000/manifest.json",
177
- "webapp_model_dir": "/data/repository_library/exports/agent_kernel/models/agentkernel_lite_100m_bitnet_15000"
178
  }
 
1
  {
2
  "artifact_kind": "agentkernel_lite_browser_bitnet_export",
3
+ "dense_dtype": "float32",
4
+ "dense_float32_include": [],
5
  "dense_tensor_count": 43,
6
+ "device": "cpu",
7
  "format": "model-stack-browser-bitnet",
8
  "layer_count": 109,
9
+ "manifest_path": "/data/agentkernel/artifacts/agentkernel_lite_encdec/recommendation_answer_v11_exact_twopass_from_v9_train_00200_lr2e6/browser_bitnet_hf/manifest.json",
10
  "max_seq_len": 1024,
11
  "model": {
12
  "activation": "silu",
13
+ "agent_policy_heads": false,
14
  "attention_bias": false,
15
+ "attn_dropout": 0.05,
16
  "attn_impl": "eager",
17
  "d_ff": 2048,
18
  "d_model": 640,
 
22
  "kv_cache_paged": true,
23
  "masking": "build_causal_mask",
24
  "max_position_embeddings": 4096,
25
+ "mlp_dropout": 0.03,
26
  "n_heads": 10,
27
  "n_kv_heads": null,
28
  "n_layers": 6,
29
  "norm": "layer",
30
  "pad_token_id": 0,
31
  "positional": "apply_rotary",
32
+ "resid_dropout": 0.03,
33
  "residual": "prenorm",
34
+ "retrieval_head_dim": null,
35
  "rms_norm_eps": 1e-06,
36
  "rope_attention_scaling": null,
37
  "rope_scaling_factor": null,
 
45
  "version": 1,
46
  "vocab_size": 8207
47
  },
48
+ "output_dir": "/data/agentkernel/artifacts/agentkernel_lite_encdec/recommendation_answer_v11_exact_twopass_from_v9_train_00200_lr2e6/browser_bitnet_hf",
49
+ "quant_exclude": [],
50
+ "quant_include": [],
51
  "quantization": {
52
  "activation_quant": "none",
53
  "modules": [
 
166
  "spin": false,
167
  "weight_opt": "none"
168
  },
169
+ "quantize_bitnet": true,
170
+ "size_bytes": 80257831,
171
+ "source_bundle_manifest_path": "/data/agentkernel/artifacts/agentkernel_lite_encdec/recommendation_answer_v11_exact_twopass_from_v9_train_00200_lr2e6/agentkernel_lite_encdec_manifest.json",
172
+ "source_model_dir": "/data/agentkernel/artifacts/agentkernel_lite_encdec/recommendation_answer_v11_exact_twopass_from_v9_train_00200_lr2e6/model",
173
+ "source_tokenizer_dir": "/data/agentkernel/artifacts/agentkernel_lite_encdec/recommendation_answer_v11_exact_twopass_from_v9_train_00200_lr2e6/tokenizer",
174
  "tokenizer": {
175
  "bos_token_id": 1,
176
  "config_path": "tokenizer/tokenizer_config.json",
 
182
  "unk_token_id": 3,
183
  "vocab_size": 8207
184
  },
185
+ "webapp_manifest_path": "/data/repository_library/exports/agent_kernel/models/agentkernel_lite_100m_bitnet_v11/manifest.json",
186
+ "webapp_model_dir": "/data/repository_library/exports/agent_kernel/models/agentkernel_lite_100m_bitnet_v11"
187
  }
dense/dec_embed_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d0b2d9d06dc96048cff8f30867bc4c520bcb7586df46a0325c95b6a93a84aeeb
3
- size 21066240
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:78c2a80005b8926af3752edc2ab4c56399688951a282a6b85bf9edbf917fe6df
3
+ size 21009920
dense/dec_norm_bias.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:bc56aaf3ca92b1288564ce350a4714e94a6021ab1a54402de1644d717312a755
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5b5e96cac42589a663a9f5682082013889507d15b2134376aa286c1c002b7fcf
3
  size 2560
dense/dec_norm_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9db5943f63cd2d6154e730c9fed0f644a1bbf7c7ed77add76e90b84c1aa4ef3b
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f3007750a249fff13b6d5325bdec87cf3e823d773e9b4ddd8661358e1c88bc4a
3
  size 2560
dense/decoder_0_cross_block_n1_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:891fb6b04974c8b901b28c09ae7ea6e5e0cdc726a71f6d17b07f1a375d4426da
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e321bd4b5815ad26ae4a2c222d73c2bd84b44c67aa3ba4a3e7425389c6eb783
3
  size 2560
dense/decoder_0_cross_block_n2_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1a719c7ed5cf7db904f9077ad94985d06a65e2d38a64340465579147ffbac8c5
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e44330e6d5b15cb95b080f739e6fe708f9915ebfc33cab76c01c6c4c2bb4e51e
3
  size 2560
dense/decoder_0_self_attn_block_attn__rope_sin.f32.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:b7e60ef23d3ccabcb925a970b51fbeaed34dbd48374a45476f3212d61fc7407f
3
- size 131072
 
 
 
 
dense/decoder_0_self_attn_block_n1_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3ecfb7e994fc3be415d9ca72f228f0a6fe1bac9ed350e75ae883ac75d4c5a431
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aa1687c4e058f3c05d691ed6f33ca18be4ec4ca67ddfe2806b29bc4926a06d87
3
  size 2560
dense/decoder_0_self_attn_block_n2_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:33d786b3efc9c3443b12cd66a09c63bd595e18ec0c9eb07fae38297f7005bf84
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aaf84ca53571564ef264cd69f05f68a5feae20d67c51111440aa38971a927f0a
3
  size 2560
dense/decoder_1_cross_block_n1_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:cbd55183838a8ab6e77bb181eeb8cf43c54e6fdec064a948915193df754f7cf9
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3ea07312d131aecfcf9aeca759f7da272db7f8669faaea13adaa03af9e74cdf7
3
  size 2560
dense/decoder_1_cross_block_n2_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5e01f19710f000878795ab523b271d8b34893f6c377f610bdd0eba9f18574a70
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a2d7b097f261522851525b00d4df45ee76aba6fd6d1c467bc592e7c75f7d0e21
3
  size 2560
dense/decoder_1_self_attn_block_attn__rope_cos.f32.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:6ad88ae23144aafc9b0fb5dbe306afcfdccb11fcf4909a512eb46125ce9f0d5b
3
- size 131072
 
 
 
 
dense/decoder_1_self_attn_block_attn__rope_sin.f32.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:b7e60ef23d3ccabcb925a970b51fbeaed34dbd48374a45476f3212d61fc7407f
3
- size 131072
 
 
 
 
dense/decoder_1_self_attn_block_n1_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3ce6ed00aee268de7d08e6e9acdecb2ce494e53740f49b397653063f65708af9
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6e04d7ae884588db5de7ef2072346f6e8b2066b26ed9f5847645221ed9539942
3
  size 2560
dense/decoder_1_self_attn_block_n2_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3aff504c8c158ad32e0d53fcde49f9c3d9d2c6d8c6ffac4793913ab8e9fc351f
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c0b1278f7154bf8b2234bdd368fbce9e7bd74601a1c793f56fa4f3586a7f7b73
3
  size 2560
dense/decoder_2_cross_block_n1_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:074691158ffb33267b6d1515d51dfccffd776d308d77ca10cd42f5c9e11d16b3
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:28376ace409fa2952ac2990043be9b9c983aea3195e64ca73a2c6ba4a7679303
3
  size 2560
dense/decoder_2_cross_block_n2_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:252019fc3c4e0a4b6e8d3b4a6612fa119688c8dda754f204055cd8276a1136ac
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1feefaae8826e4c2ad02d385f9b805519b96a7db7fe49052f1f8a8c693309339
3
  size 2560
dense/decoder_2_self_attn_block_attn__rope_cos.f32.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:6ad88ae23144aafc9b0fb5dbe306afcfdccb11fcf4909a512eb46125ce9f0d5b
3
- size 131072
 
 
 
 
dense/decoder_2_self_attn_block_attn__rope_sin.f32.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:b7e60ef23d3ccabcb925a970b51fbeaed34dbd48374a45476f3212d61fc7407f
3
- size 131072
 
 
 
 
dense/decoder_2_self_attn_block_n1_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:142d7795a7359da525adde3aa5342706ad7fddf0f79e984b0d6f95a3fb7074ec
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3731b30effc4197d021d7ed55349454c585246d030d74d9299b07aafa7b367d6
3
  size 2560
dense/decoder_2_self_attn_block_n2_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b5c5cb7840e3107062925817a43bd242b3fe2fd4a7786cc1a9460d9312c1b6eb
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:632df68307a27cfc2054c13a9b511180dce6e735c8c79dc6a420e984ffcc385c
3
  size 2560
dense/decoder_3_cross_block_n1_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:15ae5917f2aa04cc28786948c54d7375493275ebabe9e5ffc2d6b53825b7776f
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e8c667419c8df40c09c7e194573444198f3e0f8d057e61e301b293928695560
3
  size 2560
dense/decoder_3_cross_block_n2_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:caccd589d5a01abd27c8661e131ee691f8b6559a4330aeee886cb0492a73aeb3
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:770f911b6ec49b1da40240580ffac220c6308585bb36d1151a6623a5d46ef740
3
  size 2560
dense/decoder_3_self_attn_block_attn__rope_cos.f32.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:6ad88ae23144aafc9b0fb5dbe306afcfdccb11fcf4909a512eb46125ce9f0d5b
3
- size 131072
 
 
 
 
dense/decoder_3_self_attn_block_attn__rope_sin.f32.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:b7e60ef23d3ccabcb925a970b51fbeaed34dbd48374a45476f3212d61fc7407f
3
- size 131072
 
 
 
 
dense/decoder_3_self_attn_block_n1_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0ee7056ae2fdd0937969db21841862681e470303fa423918e28d426c6ea032dc
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:083fed8849ddd08ad81f0dd422aadfb0570d1f11079073bf2f85121fcdd57274
3
  size 2560
dense/decoder_3_self_attn_block_n2_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a560a1dede092c8a7dde8fa9b96baa535ef09c717dfbd8943a0d15b326d9f9d5
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:115676e9f45a24b93168c64e085aedcadfa8fe199d1a2b06b9ca67ba3615385c
3
  size 2560
dense/decoder_4_cross_block_n1_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c043209537341e6dc5f41dcf6a351fb4265565500d454fe64c667ac7b06c7e33
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b15ddc920e96f925d4bb97e943a92a38605ffd1e1f5a74373654daef985e377d
3
  size 2560
dense/decoder_4_cross_block_n2_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:598e5457b0d29193b07509b15b611a39769cdd37c70e35c9ef33339988c1541c
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:23021c9331a4ec203014d4d33f2a846ad81702b6da07a3fdcaab2ef30a4d73f6
3
  size 2560
dense/decoder_4_self_attn_block_attn__rope_cos.f32.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:6ad88ae23144aafc9b0fb5dbe306afcfdccb11fcf4909a512eb46125ce9f0d5b
3
- size 131072
 
 
 
 
dense/decoder_4_self_attn_block_attn__rope_sin.f32.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:b7e60ef23d3ccabcb925a970b51fbeaed34dbd48374a45476f3212d61fc7407f
3
- size 131072
 
 
 
 
dense/decoder_4_self_attn_block_n1_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:77f69a1f0ab7f6664172e335f1bacbede678a34d59234098b531a41f08eda87a
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f3ae1bf89ed7cdb798ec9b786ff358f126b58fb7fb6fa2b93d217a243318aeee
3
  size 2560
dense/decoder_4_self_attn_block_n2_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:2c4fcbc99f4fcf2182c7ed689ef906b7135326b75644ba9155239e6ca65c1591
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7161e760f5e51f0e019f3ddf6285c1d31819100682234328ad54c9f3fb26768c
3
  size 2560
dense/decoder_5_cross_block_n1_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:744b9f0653545a86ed4b7f963023c78c74637332d473b374e61fa19d72795ca7
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b49407ae4ad043d9c524839c8e197cfaed22ac0f9755458fc8dfb5f9501f62e5
3
  size 2560
dense/decoder_5_cross_block_n2_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:41b98dd7de3c29b4b66996542d2b1f835d8729bcd09ae96689de78daa16c2041
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d7c9bf8b3e50171e0aa8882bc82629636d6afaa7ce2b36f3612a9a79e40a5840
3
  size 2560
dense/decoder_5_self_attn_block_attn__rope_cos.f32.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:6ad88ae23144aafc9b0fb5dbe306afcfdccb11fcf4909a512eb46125ce9f0d5b
3
- size 131072
 
 
 
 
dense/decoder_5_self_attn_block_attn__rope_sin.f32.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:b7e60ef23d3ccabcb925a970b51fbeaed34dbd48374a45476f3212d61fc7407f
3
- size 131072
 
 
 
 
dense/decoder_5_self_attn_block_n1_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1fa85cc37348e322bb9b80bd0c8775dabc0e8ca898919c86bc3092effd31b7d1
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5ab4801dfa1f732a06dd0d14acfa620dbcec11568fdf8c5e333b0deb154f1a51
3
  size 2560
dense/decoder_5_self_attn_block_n2_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:2277d03a4376433872684796edfa38c6636a096944d5d8c9031b2f13bdcec68f
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c5666fab5237df5532e10c7c08ecd276dce4be357b11819655deea19d3d6b0e9
3
  size 2560
dense/enc_embed_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d0b2d9d06dc96048cff8f30867bc4c520bcb7586df46a0325c95b6a93a84aeeb
3
- size 21066240
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:78c2a80005b8926af3752edc2ab4c56399688951a282a6b85bf9edbf917fe6df
3
+ size 21009920
dense/enc_norm_bias.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0f7a17d8d5a7e35c03338862b11b135d1a0816358f766706d2f83179d4bafc40
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:73c7b6e76072acad2244afc73657d29377fd189a8af9bc41daa14aabb346606a
3
  size 2560
dense/enc_norm_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:323e132c6135d519e191f4e98eaef83861db8531b4971f7dbc32686258e791c0
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fbce78fa3910cca15c0bec29f929009612f7f5b94d1d083771ed0b114bb2a337
3
  size 2560
dense/{decoder_0_self_attn_block_attn__rope_cos.f32.bin → enc_pos_embed_weight.f32.bin} RENAMED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6ad88ae23144aafc9b0fb5dbe306afcfdccb11fcf4909a512eb46125ce9f0d5b
3
- size 131072
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:41e06124cce8055432e8e6a7a9a9a2a6e39ac672071f23dea5872bd19c51e4d9
3
+ size 10485760
dense/encoder_0_n1_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a2fca0cec706dcc674c02b5ecb68458f347cc18cbe0b5783ba08764863ca1211
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:146bdd3c5419ea957cba0919904c3fa3a18345399e1190e38b1697c5bc56a18e
3
  size 2560
dense/encoder_0_n2_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e5592bab55e8eae12dbc21470a64a4f145450f19bd28292ec789863278e122e8
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2817160dc2b42d5f8de29fb566ed4165dff109f084f8a5af1766e544a7fea509
3
  size 2560
dense/encoder_1_n1_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9efa97efc788129bea077cbe83a7d35ce84e7d90a5cdbfd87f490fbf9bd26b3d
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4b3b5cbb90b337e7c217bfe73872429a99ef7aaeda0743a01e7c78b496abce08
3
  size 2560
dense/encoder_1_n2_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:80da1c04bd2f7b3f1784986cfdd88a86f4a820b2e0c6ebd1b6b1ac4b7bb59645
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2030393e0c8cc4554a10187e03f03e541a380227d3b64c3f6455aa23d8295be6
3
  size 2560
dense/encoder_2_n1_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:73b9240c3a6ceb6a628d5b63cb7cee6f5f4f0e1506d3e59ce8422692f9afe87b
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:203d16e2f34a9fb6059e24d8bde0b3d2e85e8edbeaecb2eac7c647b9c81f1dde
3
  size 2560
dense/encoder_2_n2_weight.f32.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d4af0825ded821252546f8d1c587f59d9108b6ad5b3fec91f49de34dbf827133
3
  size 2560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a223438f7de8e65465b9c95605748e0152bc1b92d77b72d3af2a41d9ff108927
3
  size 2560