Hugging Face's logo

EdBerg
/
SmolLM2-1.7B-Instruct_Baha_1

Transformers
TensorBoard
Safetensors
Generated from Trainer
trl
sft
Model card Files Metrics Community
EdBerg commited on
Commit
8af6d04
·
verified ·
1 Parent(s): f230415

EdBerg/SmolLM2-1.7B-Instruct_Baha_1

Browse files
Files changed (18) hide show
  1. README.md +1 -1
  2. adapter_config.json +6 -6
  3. adapter_model.safetensors +2 -2
  4. runs/Feb13_01-40-05_c8fc74a74acf/events.out.tfevents.1739410806.c8fc74a74acf.465.0 +3 -0
  5. runs/Feb13_01-41-42_c8fc74a74acf/events.out.tfevents.1739410903.c8fc74a74acf.465.1 +3 -0
  6. runs/Feb13_01-43-40_c8fc74a74acf/events.out.tfevents.1739411020.c8fc74a74acf.2443.0 +3 -0
  7. runs/Feb13_01-49-00_c8fc74a74acf/events.out.tfevents.1739411342.c8fc74a74acf.2443.1 +3 -0
  8. runs/Feb13_01-49-36_c8fc74a74acf/events.out.tfevents.1739411377.c8fc74a74acf.2443.2 +3 -0
  9. runs/Feb13_01-49-53_c8fc74a74acf/events.out.tfevents.1739411394.c8fc74a74acf.2443.3 +3 -0
  10. runs/Feb13_01-50-07_c8fc74a74acf/events.out.tfevents.1739411408.c8fc74a74acf.2443.4 +3 -0
  11. runs/Feb13_01-56-47_c8fc74a74acf/events.out.tfevents.1739411808.c8fc74a74acf.2443.5 +3 -0
  12. runs/Feb13_01-58-26_c8fc74a74acf/events.out.tfevents.1739411907.c8fc74a74acf.2443.6 +3 -0
  13. runs/Feb13_01-58-44_c8fc74a74acf/events.out.tfevents.1739411924.c8fc74a74acf.2443.7 +3 -0
  14. runs/Feb13_01-58-57_c8fc74a74acf/events.out.tfevents.1739411937.c8fc74a74acf.2443.8 +3 -0
  15. runs/Feb13_01-59-53_c8fc74a74acf/events.out.tfevents.1739411994.c8fc74a74acf.2443.9 +3 -0
  16. runs/Feb13_03-22-54_c8fc74a74acf/events.out.tfevents.1739416974.c8fc74a74acf.2443.10 +3 -0
  17. tokenizer.json +1 -6
  18. training_args.bin +1 -1
README.md CHANGED
@@ -27,7 +27,7 @@ print(output["generated_text"])
27
 
28
  ## Training procedure
29
 
30
- [<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/harpermia882/huggingface/runs/9z7s33ce)
31
 
32
  This model was trained with SFT.
33
 
 
27
 
28
  ## Training procedure
29
 
30
+ [<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/harpermia882/huggingface/runs/4kdpchsk)
31
 
32
  This model was trained with SFT.
33
 
adapter_config.json CHANGED
@@ -13,23 +13,23 @@
13
  "layers_pattern": null,
14
  "layers_to_transform": null,
15
  "loftq_config": {},
16
- "lora_alpha": 32,
17
  "lora_bias": false,
18
- "lora_dropout": 0.05,
19
  "megatron_config": null,
20
  "megatron_core": "megatron.core",
21
  "modules_to_save": null,
22
  "peft_type": "LORA",
23
- "r": 32,
24
  "rank_pattern": {},
25
  "revision": null,
26
  "target_modules": [
 
27
  "gate_proj",
28
- "up_proj",
29
  "k_proj",
30
  "v_proj",
31
- "q_proj",
32
- "o_proj",
33
  "down_proj"
34
  ],
35
  "task_type": "CAUSAL_LM",
 
13
  "layers_pattern": null,
14
  "layers_to_transform": null,
15
  "loftq_config": {},
16
+ "lora_alpha": 8,
17
  "lora_bias": false,
18
+ "lora_dropout": 0.0,
19
  "megatron_config": null,
20
  "megatron_core": "megatron.core",
21
  "modules_to_save": null,
22
  "peft_type": "LORA",
23
+ "r": 8,
24
  "rank_pattern": {},
25
  "revision": null,
26
  "target_modules": [
27
+ "q_proj",
28
  "gate_proj",
29
+ "o_proj",
30
  "k_proj",
31
  "v_proj",
32
+ "up_proj",
 
33
  "down_proj"
34
  ],
35
  "task_type": "CAUSAL_LM",
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:75cb2c847ea9a42324fe42149ea06c5b7d321001f588e312128998b4af2b5235
3
- size 144748392
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6abf08d52d0b16874c6e34ca8a2db14d2f78e355cd3fb43d58d2987260673905
3
+ size 36220072
runs/Feb13_01-40-05_c8fc74a74acf/events.out.tfevents.1739410806.c8fc74a74acf.465.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:52d5bbb6d5d4277a29483cc4c2720db1985b3b398e5383b433331004dcd2acd9
3
+ size 6130
runs/Feb13_01-41-42_c8fc74a74acf/events.out.tfevents.1739410903.c8fc74a74acf.465.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8c0e92f9758a2772a06b68bb46fa713a62ec0f9a0275245bb4d778f7b846891b
3
+ size 4184
runs/Feb13_01-43-40_c8fc74a74acf/events.out.tfevents.1739411020.c8fc74a74acf.2443.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fac3a69138d1e729c16e82bcd5a92128e777cb7b4906256fa816345745abeeae
3
+ size 6130
runs/Feb13_01-49-00_c8fc74a74acf/events.out.tfevents.1739411342.c8fc74a74acf.2443.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:24a2f9735ac343674e4de0d6121d28d0582e468fd61cfbd757d5e64caadd5ba9
3
+ size 6130
runs/Feb13_01-49-36_c8fc74a74acf/events.out.tfevents.1739411377.c8fc74a74acf.2443.2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c10a50d826856eef764fcbf61aeca7e9af316f7eee8394937479e23b2c552efd
3
+ size 6130
runs/Feb13_01-49-53_c8fc74a74acf/events.out.tfevents.1739411394.c8fc74a74acf.2443.3 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:396fea64c7a4f292c6de94ce618b46f54e119caf379eee2bae98c5b43518bd1f
3
+ size 6130
runs/Feb13_01-50-07_c8fc74a74acf/events.out.tfevents.1739411408.c8fc74a74acf.2443.4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:849573809ba45ff7d188a2ab642f4570a50d46bb23a912efe5c6f65936aab04d
3
+ size 20620
runs/Feb13_01-56-47_c8fc74a74acf/events.out.tfevents.1739411808.c8fc74a74acf.2443.5 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:acfdb9c8ea1f1725f7b6e9e1ada34c83a7599ef835a66551d4b0c89695464c04
3
+ size 6337
runs/Feb13_01-58-26_c8fc74a74acf/events.out.tfevents.1739411907.c8fc74a74acf.2443.6 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:33a02448dcca27c6aea56851cbc889abf8f010cbb57f91de5634003e8926ce06
3
+ size 6337
runs/Feb13_01-58-44_c8fc74a74acf/events.out.tfevents.1739411924.c8fc74a74acf.2443.7 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7d04a1b04b8ba0c4d294af9ebd3918d7d9a856e830629ceb71625b20ad9b4386
3
+ size 7165
runs/Feb13_01-58-57_c8fc74a74acf/events.out.tfevents.1739411937.c8fc74a74acf.2443.8 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b6e061317b84e2f04c110eac32108a297e623564126ac7840e5a995d18b646d4
3
+ size 32212
runs/Feb13_01-59-53_c8fc74a74acf/events.out.tfevents.1739411994.c8fc74a74acf.2443.9 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:139b2638d743f1d333dc262bd6b6dbb726a7556c9468056b57bdf6753d071c80
3
+ size 849976
runs/Feb13_03-22-54_c8fc74a74acf/events.out.tfevents.1739416974.c8fc74a74acf.2443.10 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8eb26a4f6d1f8cec1ef26b7f05464fc99233a338a111022710437708a3e7cb47
3
+ size 1271976
tokenizer.json CHANGED
@@ -1,11 +1,6 @@
1
  {
2
  "version": "1.0",
3
- "truncation": {
4
- "direction": "Right",
5
- "max_length": 1024,
6
- "strategy": "LongestFirst",
7
- "stride": 0
8
- },
9
  "padding": null,
10
  "added_tokens": [
11
  {
 
1
  {
2
  "version": "1.0",
3
+ "truncation": null,
 
 
 
 
 
4
  "padding": null,
5
  "added_tokens": [
6
  {
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:12cd2b0fd4c5f2ec0aa665ea6d6d239d5b27e06098f0bd8f491cd87138d52d93
3
  size 5624
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:104a4a0f1675518101c2be1f8a1b916ae0e6573b73c5556dd411a17374deca39
3
  size 5624